Best AI subscriptions for your money

Which AI is worth paying for?

Plans ranked by the model work a dollar buys, counting your time.

Value per dollar = what a month of the plan’s tokens would cost on the API, divided by its price plus the hours you spend using them at $50 an hour. A token is worth the API price of the cheapest model our ranking rates at least as high. Plans that publish no convertible allowance, free plans and non-comparable offers have no figure.

PlanValue per dollarPriceBest use of the allowanceVerdict
Claude Ultra1.2×USD 200/moClaude Opus 5.5, 29B tokens a month, 158 h of your timeTen times Pro's weekly Claude usage without a five-hour limit once the account is verified, plus a separate monthly API budget; credits require claiming and cannot extend interactive plan limits.
Claude Pro1.2×USD 20/moClaude Opus 5.5, 2.9B tokens a month, 16 h of your timeGenerous Opus allowance for regular chat and coding in one subscription.
Claude Max1.2×USD 100/moClaude Opus 5.5, 14B tokens a month, 77 h of your timeHigher Claude usage plus a separate monthly API budget for custom apps and agents; credits require claiming and cannot extend interactive plan limits.
ChatGPT Pro 2001.1×USD 200/moGPT-6 Astra, 3.9B tokens a month, 21 h of your timeMore agent headroom without a five-hour limit; OpenAI canceled the September reduction, so every Pro 200 subscription has the higher allowance again.
ChatGPT Pro 5001.1×USD 500/moGPT-6 Astra, 9.5B tokens a month, 51 h of your timeThe largest of the three Pro allowances and the only Pro tier with Ultrafast; per dollar it buys about what Pro 200 does, so pay for it when Pro 200 runs out or Ultrafast speed matters.
ChatGPT Pro 1001.1×USD 100/moGPT-6 Astra, 1.8B tokens a month, 9.8 h of your timeMore agent capacity without a five-hour limit, with the broader ChatGPT toolset.
GitHub Copilot Max0.78×USD 100/moClaude Opus 5.5, 577M tokens a month, 3.1 h of your timeA sizeable published coding budget for people who will use the extra credits.
ChatGPT Plus0.76×USD 20/moGPT-6 Astra, 106M tokens a month, 34 min of your timeA strong all-round bundle; sustained agent work can outgrow its limits.
GitHub Copilot Pro+0.75×USD 39/moClaude Opus 5.5, 202M tokens a month, 1.1 h of your timePremium model choice and a larger published credit pool for regular coding.
Venice Max0.60×USD 200/moClaude Opus 5.5, 649M tokens a month, 3.5 h of your timeMore premium capacity, worth the bill only when regularly consumed.
Venice Pro Plus0.59×USD 68/moClaude Opus 5.5, 216M tokens a month, 1.2 h of your timeA much larger premium pool, with separate model classes and limited rollover.
Kilo / Kilo Pass Pass Expert0.55×USD 208.95/moClaude Opus 5.5, 574M tokens a month, 3.1 h of your timeA large prepaid balance, without guaranteed recurring promotional uplift.
Kilo / Kilo Pass Pass Pro0.55×USD 51.45/moClaude Opus 5.5, 141M tokens a month, 46 min of your timeMore prepaid model access, with separate cloud costs and expiring bonus credit.
Kilo / Kilo Pass Pass Starter0.55×USD 19.95/moClaude Opus 5.5, 55M tokens a month, 18 min of your timeBroad model routing with prepaid credit; recurring bonuses are conditional.
OpenCode Go0.45×USD 10/moGemini 4 Argon, 87M tokens a month, 28 min of your timeLow-cost coding access with explicit model-dependent token pools; rolling limits, context prices and contributor terms affect usable capacity.
TRAE Ultra0.39×USD 200/moGPT-6.1 Sol, 1.2B tokens a month, 6.2 h of your timeThe largest published TRAE budget, but at face value: $200 of usage for $200, plus early model access.
TRAE Pro+0.39×USD 60/moGPT-6.1 Sol, 346M tokens a month, 1.9 h of your timeThree times Pro’s budget for three times the price; the extra concurrency is the real difference.
TRAE Pro0.39×USD 20/moGPT-6.1 Sol, 115M tokens a month, 37 min of your timeThe included usage now equals the price, so the plan buys concurrency and unlimited autocomplete rather than a discount.
Zed Pro0.34×USD 10/moClaude Opus 5.5, 13M tokens a month, 4 min of your timeA useful editor subscription, but only $5 of hosted model usage is included.
GitHub Copilot Pro0.10×USD 10/moGPT-6.1 Terra, 173M tokens a month, 56 min of your timeAn affordable editor companion with a finite model-dependent credit budget.
Meta Muse Code High Usage0.059×USD 15/moMuse Spark 1.3, 4.0B tokens a month, 22 h of your timeMore included coding usage at a modest price; five-hour resets and CLI-only access remain.
Meta Muse Code Power Usage0.059×USD 50/moMuse Spark 1.3, 13B tokens a month, 71 h of your timeAdds the distinct Muse Wonder Full model and substantially more usage for sustained coding.
Mistral Vibe / Le Chat Pro0.058×USD 14.99/moDeepSeek V4.1 Flash (1017), 1.5B tokens a month, 8.0 h of your timeChat, coding and $25.50 monthly API credit in one bundle; fair-use limits remain and the API budget is lower than before.
Venice Pro0.053×USD 18/moClaude Opus 5.5, 2.9M tokens a month, 1 min of your timeBroad text access, but the frontier-model premium budget is only $1 equivalent.
Kimi Ultra0.048×USD 199/moKimi K3, 3.1B tokens a month, 17 h of your timeTen times the entry Agent quota for sustained use; the total token grant remains unpublished.
Kimi Max0.048×USD 99/moKimi K3, 1.5B tokens a month, 8.1 h of your timeFive times the entry Agent quota, with a higher recurring commitment.
Kimi Pro0.046×USD 39/moKimi K3, 483M tokens a month, 2.6 h of your timeTwice the entry Agent quota and access to longer-context and HighSpeed coding.
Z.ai Max0.046×USD 168/moGLM-5.3, 6.3B tokens a month, 34 h of your timeA large published coding allocation whose value depends on using it.
Z.ai Pro0.045×USD 80/moGLM-5.3, 2.7B tokens a month, 15 h of your timeMore coding credits, with two simultaneous caps and separate client requirements.
Z.ai Lite0.044×USD 18/moGLM-5.3, 449M tokens a month, 2.4 h of your timeA focused coding subscription with explicit short-window and weekly caps.
Kimi Plus0.036×USD 19/moKimi K3, 110M tokens a month, 36 min of your timeChat and coding share a modest monthly pool; the new plan removes the legacy weekly coding cap.
v0 Plus0.033×USD 30/moClaude Opus 4.8, 133M tokens a month, 43 min of your timeApp-building credit and daily login grants at a moderate recurring price.
Alibaba Cloud Coding Plan Pro0.024×USD 50/moQwen 3.7 Plus, 14B tokens a month, 74 h of your timeA substantial published call quota, offset by limited purchase slots and client restrictions.
v0 Business0.016×USD 100/moClaude Opus 4.8, 133M tokens a month, 43 min of your timeThe same included inference as Plus; the premium mainly buys business controls.
JetBrains AI Ultimate0.015×USD 30/moClaude Opus 5.5, 2.7M tokens a month, 47 h of your timeShared AI credits for IDE chat and agents; the value figure uses the verified AI Assistant chat model roster.
JetBrains AI Pro0.015×USD 10/moClaude Opus 5.5, 758k tokens a month, 13 h of your timeShared AI credits for IDE chat and agents; the value figure uses the verified AI Assistant chat model roster.
Monica Pro<0.01×USD 9.90/moGPT-6.1 Sol, 470k tokens a month, 8.3 h of your timeA low annual entry price, with separate query and advanced-credit limits.
Monica Ultra<0.01×USD 99.90/moGPT-6.1 Sol, 3.8M tokens a month, 66 h of your timeMore advanced credits, but a large annual commitment and qualified unlimited claims.
Kagi Assistant Starter<0.01×USD 5/moDeepSeek V4.1 Flash, 705k tokens a month, 12 h of your timeLow-cost search and metered Assistant access now include Haiku 5.5; the largest models still require Ultimate.

Open the ranking for each plan’s allowance, arithmetic, sources and the full method.

Compare AI subscriptions for everyday use, research, coding and work. 178 offers and configurations across 45 services, including regional, team and unresolved offers. Individual USD recommendations are a smaller editorial shortlist.

Daily everyday use, up to $20/month

Editorial recommendations: start with entry paid plans for daily use, free options for occasional use, and intermediate upgrades for heavy use. The order is a starting point, not a measured task-success ranking. Quotas depend on workload.

PlanMonthly USDWhy consider itAllowanceSources
ChatGPT Plus$20One entry-level paid plan for both everyday chat and focused agent work.Estimated local messages per 5 hours: GPT-6 Astra 5–45, GPT-6.1 Sol 15–160, GPT-6.1 Terra 35–285, GPT-6 Sol 15–150. Agent usage is shared. Task complexity and weekly limits affect capacity. GPT-5.5 retired from ChatGPT and Codex on October 14; Codex with ChatGPT stopped serving it at 19:00 UTC. API-key access is unaffected (official pricing and changelog, October 15).Checked 2026-11-02. OpenAI plans and usage
Claude Pro$20A starting point for work that moves between documents, chat and Claude Code.Variable usage; shared chat and coding limits. Chat and coding share usage. Absolute monthly capacity is undisclosed. With Opus 5.5 (anthropic.com, September 22, 2026) Anthropic raised five-hour limits on paid plans by an unpublished amount and granted a one-off limit reset; the announcement names no weekly-limit change.Checked 2026-11-02. Claude plans and access · Claude shared usage limits
Google AI Pro$19.99Consider it when you already work in Gmail and Docs and want AI in those tools.Gemini app: 4× standard limits. Antigravity: shared work-dependent five-hour and weekly limits. App and coding quotas differ. Paid non-trial Pro includes Argon and premium Claude in the Antigravity app, CLI, standalone IDE and extensions; trial access differs.Checked 2026-11-01. Google US recurring price (offer terms) · Gemini app access and limits · Antigravity coding access and limits · Gemini image models and plan access
ChatGPT Go$8A small recurring commitment for light use.Lightweight Codex use; no fixed allowance stated here. Agent usage is shared. Task complexity and weekly limits affect capacity. GPT-5.5 retired from ChatGPT and Codex on October 14; Codex with ChatGPT stopped serving it at 19:00 UTC. API-key access is unaffected (official pricing and changelog, October 15).Checked 2026-11-02. OpenAI plans and usage
ChatGPT Free$0Start here to try chat and coding before committing to a bill.Limited access for quick tasks. Agent usage is shared. Task complexity and weekly limits affect capacity. GPT-5.5 retired from ChatGPT and Codex on October 14; Codex with ChatGPT stopped serving it at 19:00 UTC. API-key access is unaffected (official pricing and changelog, October 15).Checked 2026-11-02. OpenAI plans and usage
Claude Free$0Try writing, analysis and conversation without a subscription.Limited, variable usage. Chat and coding share usage. Absolute monthly capacity is undisclosed. With Opus 5.5 (anthropic.com, September 22, 2026) Anthropic raised five-hour limits on paid plans by an unpublished amount and granted a one-off limit reset; the announcement names no weekly-limit change.Checked 2026-11-02. Claude plans and access · Claude shared usage limits
Gemini Free$0A free starting point for people who prefer the Gemini app.Standard Gemini app limits. Feature and model limits can change.Checked 2026-11-01. Gemini app access and limits
X Plus$8One shared weekly allowance for chat and coding with Grok.Basic shared weekly Grok usage. Absolute task and token counts are not published. Grok models only; cloud agents, automations, Grok Bot and third-party model access require Premium or above. One weekly Grok usage pool is shared across Grok and Cursor. On Premium and above, third-party models in Cursor are billed separately from the first use; they are not part of the included allowance. Taxes and fees are extra. Switching from a legacy plan is optional and permanent. The October 28 Grok Bot FAQ confirms Bot is included on Premium, Super and Ultra, using the same weekly pool; Plus and Free exclude it. Bot selects models automatically without a model picker. For legacy accounts with linked Cursor and SuperGrok subscriptions, Bot uses the larger allowance rather than adding them, and both subscriptions continue billing separately.Checked 2026-11-04. The X Subscription pricing and access · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation
Mistral Vibe Pro$14.99A lower-priced chat and coding bundle if Mistral meets your needs.Higher chat/coding limits; $25.50 monthly API credit Web/mobile chat and CLI/IDE coding are bundled, with fair-use caps. Regular Pro includes $25.50 monthly API credit; eligible education Pro has $15 instead. Team prices are per seat and the minimum purchase still needs verification. The Studio/API model catalogue offers GLM-5.3; GLM-5.2 was retired on October 31 (deprecated and retired models table, read November 1). Mistral Large 4 launched in public API preview October 6, with weights promised for later in the month. Included API usage is shared with Studio and Vibe; the preview announcement does not establish Le Chat access. Large 4 now has an exact ranked row and is included in the API access roster (official docs and launch announcement, read 2026-10-08). October 29: the API catalogue includes unmodified DeepSeek V4.1 Flash 1017 at $0.30 input, $0.007 cached input and $1.20 output per million tokens. This is an alternative use of the existing $25.50 regular Pro or $15 education API credit, not evidence of Le Chat access or an extra pool. October 30: Mistral released the improved Large 4 weights on October 29; the API model mistral-large-4 now costs $0.70 input, $0.07 cached input and $2.10 output per million tokens (launch post, API pricing table and model card, whose version label remains 26.10). Large 4 stays one alternative use of the same $25.50 regular Pro or $15 education API credit. October 31: Mistral deprecated Mistral Medium 3.5 and Mistral Large 3 (deprecation October 31, 2026; retirement April 30, 2027; replacement Mistral Large 4) and removed both from the API pricing table; they stay usable until retirement but no longer have a published rate. DeepSeek V4.1 Flash 1017 now also accepts image input at unchanged rates. The improved October 29 Mistral Large 4 (ranking row mistral-large-4-oss, High reasoning by default) is available through the Mistral API and Vibe Code, so it is one more alternative use of the same API credit, not an extra pool.Checked 2026-10-31. Mistral Vibe / Le Chat pricing and access · Mistral Vibe / Le Chat usage documentation · Mistral Vibe / Le Chat usage documentation · Mistral Vibe / Le Chat usage documentation · Mistral Vibe / Le Chat usage documentation
Perplexity Pro$20Consider it when cited search and research matter most.Search limits; no monthly Computer credits; weekly Light-effort Computer usage of unpublished size Nano Banana 2.1 generates and edits images inside Computer for subscribers with Computer access. The October 19 release note establishes image-tool access, without publishing an image allowance or model credit rate (read October 25). Ask/search and Computer use different allowances. Pro has no recurring Computer credit allocation; Max has 10,000/month. Consumer search/research caps are not fully numeric in the current comparison table. Ignore one-time credit bonuses. Paid consumer Pro plans now also include a weekly Computer usage allowance of unpublished size; without any available credits it runs tasks only on the Light effort setting, where Perplexity picks the model, and it does not roll over. Promotional Pro plans, Max and Enterprise do not get it (perplexity.ai credits help, reread 2026-10-29). The October 27 Search model table includes GPT-6.1 Sol, Gemini 3.8 Flash, Sonnet 5.5, Kimi K3, GLM-5.3, Grok 4.7 and Nemotron 3.5 Super VL; Max adds GPT-6 Astra, Gemini 4 Argon, Opus 5.5 and Fable 5.1. Search, Council, Comet and Computer use separate selectors. Computer Custom can select a coordinating model and effort when spending credits; supporting agents remain independently selected. Credits cost $1 per 100 to purchase, but model consumption rates are unpublished. One-time 4,000/35,000-credit bonuses are excluded from recurring capacity (official help, October 21). October 28 review: Max can select Ultrafast GPT-6 Astra or GPT-6.1 Sol in Computer on the web; it consumes more credits than Normal or Fast, with no published conversion. Some Free accounts receive one-time Computer turns, without a recurring refill and with one agent at a time. Current consumer prices and the October 27 Search model table remain unchanged; the credit guide was last modified October 27. The API remains separately billed. October 29: included Pro Computer usage is rolling out gradually; the account Usage page establishes availability. With credits, other effort settings including Custom are available and higher effort consumes the weekly allowance faster. The weekly quantity and model rates remain unpublished.Checked 2026-10-28. Perplexity pricing and access · Perplexity usage documentation · Perplexity usage documentation · Perplexity usage documentation · Perplexity usage documentation · Perplexity usage documentation · Perplexity usage documentation · Perplexity usage documentation
Abacus ChatLLM$10A low-priced way to access several model families in one app.20,000 credits/month; 3 AI Agent tasks and 3 Studio conversations (up to 2,500 credits each); chat token allowance unpublished First-month discount excluded; vendor capacity claims untested. ChatLLM Basic is $10/user/month after the introductory month, with20,000 monthly credits. Pro is $20/user/month with30,000 credits; Pro remains a catalogue candidate. Basic has three AI Agent tasks monthly; Studio has three conversations capped at2,500credits each. Pro removes those task/conversation limits, but usage still consumes credits. Official FAQ states credits are not tokens and gives only up-to token claims; purchasing10,000credits for$10 does not establish an API-dollar conversion. Advanced models and features share credits, and model fallback/attachment limits apply. Read October24.Checked 2026-10-24. Abacus ChatLLM pricing and access · Abacus ChatLLM usage documentation
Google AI Plus$4.99A low-cost upgrade when free Gemini is the main limit.2x Free Gemini usage; 200 Flow credits/month Nano Banana 2.1 image generation and editing is included in the Gemini app on Google AI Plus, Pro and Ultra. Its image allowance is separate from text usage; Google does not publish a 2.1-specific image count on the access page (official US image-generation page, read October 25). Gemini uses compute-based five-hour and weekly controls; Flow credits are separate. US pricing and country eligibility apply. Gemini 4 Argon is included in AI Pro and Ultra in the Gemini app and Search. Paid Pro, eligible premium-device Pro trials and Ultra include Argon in the Antigravity app, CLI, standalone IDE and IDE extensions (official model matrix, October 27). Paid non-trial Pro and Ultra include Claude Opus 5.5 and Sonnet 5.5 thinking in Antigravity. Free and Plus do not. Gemini 3.8 Flash is available across Antigravity tiers. Older Opus/Sonnet 4.6 and GPT-OSS access there ends November 18 (official model matrix reviewed 2026-11-03). Gemini CLI and the Code Assist IDE extensions stopped serving free individual, Google AI Pro and Ultra requests on June 18, 2026. Consumer coding now uses Antigravity CLI, app and IDE extensions. Its included quota is work-dependent, with five-hour and weekly controls for paid plans; the old Gemini CLI request counts do not apply. Enterprise Code Assist and paid API-key access are separate.Checked 2026-10-31. Google AI / Gemini pricing and access · Google AI / Gemini usage documentation · Google AI / Gemini usage documentation · Google AI / Gemini usage documentation · Google AI / Gemini usage documentation

Editorial shortlist

Reference prices, including offers outside the default profile. Offers older than 30 days leave the recommendations.

PlanMonthly USDWhy consider itAllowanceSources
X Plus$8One shared weekly allowance for chat and coding with Grok.Basic shared weekly Grok usage. Absolute task and token counts are not published. Grok models only; cloud agents, automations, Grok Bot and third-party model access require Premium or above. One weekly Grok usage pool is shared across Grok and Cursor. On Premium and above, third-party models in Cursor are billed separately from the first use; they are not part of the included allowance. Taxes and fees are extra. Switching from a legacy plan is optional and permanent. The October 28 Grok Bot FAQ confirms Bot is included on Premium, Super and Ultra, using the same weekly pool; Plus and Free exclude it. Bot selects models automatically without a model picker. For legacy accounts with linked Cursor and SuperGrok subscriptions, Bot uses the larger allowance rather than adding them, and both subscriptions continue billing separately.Checked 2026-11-04. The X Subscription pricing and access · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation
X Premium$30One shared weekly allowance for chat and coding with Grok.4× Plus shared weekly Grok usage. Absolute task and token counts are not published. Includes the latest Grok models, Grok Bot, Cursor cloud agents, automations and code reviews; third-party usage costs extra. One weekly Grok usage pool is shared across Grok and Cursor. On Premium and above, third-party models in Cursor are billed separately from the first use; they are not part of the included allowance. Taxes and fees are extra. Switching from a legacy plan is optional and permanent. The October 28 Grok Bot FAQ confirms Bot is included on Premium, Super and Ultra, using the same weekly pool; Plus and Free exclude it. Bot selects models automatically without a model picker. For legacy accounts with linked Cursor and SuperGrok subscriptions, Bot uses the larger allowance rather than adding them, and both subscriptions continue billing separately.Checked 2026-11-04. The X Subscription pricing and access · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation
X Super$100One shared weekly allowance for chat and coding with Grok.15× Plus shared weekly Grok usage. Absolute task and token counts are not published. Includes the latest Grok models, Grok Bot, Cursor cloud agents, automations and code reviews; third-party usage costs extra. One weekly Grok usage pool is shared across Grok and Cursor. On Premium and above, third-party models in Cursor are billed separately from the first use; they are not part of the included allowance. Taxes and fees are extra. Switching from a legacy plan is optional and permanent. The October 28 Grok Bot FAQ confirms Bot is included on Premium, Super and Ultra, using the same weekly pool; Plus and Free exclude it. Bot selects models automatically without a model picker. For legacy accounts with linked Cursor and SuperGrok subscriptions, Bot uses the larger allowance rather than adding them, and both subscriptions continue billing separately.Checked 2026-11-04. The X Subscription pricing and access · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation
X Ultra$200One shared weekly allowance for chat and coding with Grok.50× Plus shared weekly Grok usage. Absolute task and token counts are not published. Includes the latest Grok models, Grok Bot, Cursor cloud agents, automations and code reviews; third-party usage costs extra. One weekly Grok usage pool is shared across Grok and Cursor. On Premium and above, third-party models in Cursor are billed separately from the first use; they are not part of the included allowance. Taxes and fees are extra. Switching from a legacy plan is optional and permanent. The October 28 Grok Bot FAQ confirms Bot is included on Premium, Super and Ultra, using the same weekly pool; Plus and Free exclude it. Bot selects models automatically without a model picker. For legacy accounts with linked Cursor and SuperGrok subscriptions, Bot uses the larger allowance rather than adding them, and both subscriptions continue billing separately.Checked 2026-11-04. The X Subscription pricing and access · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation · The X Subscription usage documentation
ChatGPT Free$0Start here to try chat and coding before committing to a bill.Limited access for quick tasks. Agent usage is shared. Task complexity and weekly limits affect capacity. GPT-5.5 retired from ChatGPT and Codex on October 14; Codex with ChatGPT stopped serving it at 19:00 UTC. API-key access is unaffected (official pricing and changelog, October 15).Checked 2026-11-02. OpenAI plans and usage
ChatGPT Go$8A small recurring commitment for light use.Lightweight Codex use; no fixed allowance stated here. Agent usage is shared. Task complexity and weekly limits affect capacity. GPT-5.5 retired from ChatGPT and Codex on October 14; Codex with ChatGPT stopped serving it at 19:00 UTC. API-key access is unaffected (official pricing and changelog, October 15).Checked 2026-11-02. OpenAI plans and usage
ChatGPT Plus$20One entry-level paid plan for both everyday chat and focused agent work.Estimated local messages per 5 hours: GPT-6 Astra 5–45, GPT-6.1 Sol 15–160, GPT-6.1 Terra 35–285, GPT-6 Sol 15–150. Agent usage is shared. Task complexity and weekly limits affect capacity. GPT-5.5 retired from ChatGPT and Codex on October 14; Codex with ChatGPT stopped serving it at 19:00 UTC. API-key access is unaffected (official pricing and changelog, October 15).Checked 2026-11-02. OpenAI plans and usage
ChatGPT Pro 100$100The first higher-usage step for someone already constrained by Plus.More usage than Plus with no five-hour limit; OpenAI doubled the Work and Codex allowance at the end of October; weekly limits can apply; no per-model estimates are published for Pro. Includes $100/month of eligible API Platform credit after activation. Agent usage is shared. Task complexity and weekly limits affect capacity. GPT-5.5 retired from ChatGPT and Codex on October 14; Codex with ChatGPT stopped serving it at 19:00 UTC. API-key access is unaffected (official pricing and changelog, October 15). Paid Pro accounts active for at least seven days can link an API organization they own using the same email to claim monthly API credits. The first grant is prorated; unused credit expires each billing cycle. API organization/model eligibility and rate limits apply. Credits cover eligible API Platform usage, excluding Codex with API keys, and do not increase signed-in Work or Codex limits.Checked 2026-11-02. OpenAI plans and usage · ChatGPT Pro tiers and allowance changes · ChatGPT Pro monthly API credits
ChatGPT Pro 200$200Consider this tier when sustained agent work exhausts the cheaper Pro plan.More usage than Pro 100 with no five-hour limit; weekly limits can apply; tasks remain metered. Includes $200/month of eligible API Platform credit after activation. Agent usage is shared. Task complexity and weekly limits affect capacity. GPT-5.5 retired from ChatGPT and Codex on October 14; Codex with ChatGPT stopped serving it at 19:00 UTC. API-key access is unaffected (official pricing and changelog, October 15). OpenAI canceled the Pro 200 reduction announced on September 29: new and existing Pro 200 subscriptions now get the higher Work and Codex allowance, and lower allowances doubled to match (help.openai.com, read 2026-11-01). Paid Pro accounts active for at least seven days can link an API organization they own using the same email to claim monthly API credits. The first grant is prorated; unused credit expires each billing cycle. API organization/model eligibility and rate limits apply. Credits cover eligible API Platform usage, excluding Codex with API keys, and do not increase signed-in Work or Codex limits.Checked 2026-11-02. OpenAI plans and usage · ChatGPT Pro tiers and allowance changes · ChatGPT Pro monthly API credits
Claude Free$0Try writing, analysis and conversation without a subscription.Limited, variable usage. Chat and coding share usage. Absolute monthly capacity is undisclosed. With Opus 5.5 (anthropic.com, September 22, 2026) Anthropic raised five-hour limits on paid plans by an unpublished amount and granted a one-off limit reset; the announcement names no weekly-limit change.Checked 2026-11-02. Claude plans and access · Claude shared usage limits
Claude Pro$20A starting point for work that moves between documents, chat and Claude Code.Variable usage; shared chat and coding limits. Chat and coding share usage. Absolute monthly capacity is undisclosed. With Opus 5.5 (anthropic.com, September 22, 2026) Anthropic raised five-hour limits on paid plans by an unpublished amount and granted a one-off limit reset; the announcement names no weekly-limit change.Checked 2026-11-02. Claude plans and access · Claude shared usage limits
Claude Max$100The first upgrade to consider if Claude is your main working environment.5× Pro usage every five hours and every week, plus $100/month Claude API credit. Chat and coding share usage. Absolute monthly capacity is undisclosed. With Opus 5.5 (anthropic.com, September 22, 2026) Anthropic raised five-hour limits on paid plans by an unpublished amount and granted a one-off limit reset; the announcement names no weekly-limit change. Monthly API credits are rolling out, claimable after seven days on Max into one linked Console organization. They expire each billing cycle and cannot pay for interactive Claude Code or extra subscription usage.Checked 2026-11-02. Claude plans and access · Claude shared usage limits · Claude Max plan · Claude Max and Ultra monthly API credits
Claude Ultra$200Extra capacity for sustained work in Claude and Claude Code.10× Pro usage every week with no five-hour limit for verified accounts, plus $200/month Claude API credit. Chat and coding share usage. Absolute monthly capacity is undisclosed. With Opus 5.5 (anthropic.com, September 22, 2026) Anthropic raised five-hour limits on paid plans by an unpublished amount and granted a one-off limit reset; the announcement names no weekly-limit change. Formerly Max 20x: existing subscribers moved to Ultra automatically at the same price (support.claude.com, read 2026-10-31). Until Anthropic verifies the account automatically, a temporary limit replaces the missing five-hour window. Monthly API credits are rolling out, claimable after seven days on Ultra into one linked Console organization. They expire each billing cycle and cannot pay for interactive Claude Code or extra subscription usage.Checked 2026-11-02. Claude plans and access · Claude shared usage limits · Claude Ultra plan · Claude Max and Ultra monthly API credits
Gemini Free$0A free starting point for people who prefer the Gemini app.Standard Gemini app limits. Feature and model limits can change.Checked 2026-11-01. Gemini app access and limits
Google AI Pro$19.99Consider it when you already work in Gmail and Docs and want AI in those tools.Gemini app: 4× standard limits. Antigravity: shared work-dependent five-hour and weekly limits. App and coding quotas differ. Paid non-trial Pro includes Argon and premium Claude in the Antigravity app, CLI, standalone IDE and extensions; trial access differs.Checked 2026-11-01. Google US recurring price (offer terms) · Gemini app access and limits · Antigravity coding access and limits · Gemini image models and plan access
Copilot Free$0A free way to try code suggestions in your existing editor.2,000 completions/month; limited AI credits. Credits are not tasks. Models consume them at different rates.Checked 2026-11-01. GitHub Copilot plans · GitHub credit billing
Copilot Pro$10A low-cost option for code suggestions and lighter IDE assistance.1,500 AI credits/month including flex allotment. Credits are not tasks. Models consume them at different rates.Checked 2026-11-01. GitHub Copilot plans · GitHub credit billing
Copilot Pro+$39Consider it for premium model choices inside a familiar IDE.7,000 AI credits/month including flex allotment. Credits are not tasks. Models consume them at different rates. Fable 5.1 retains prompts and outputs for safety classifiers by default; an enterprise ZDR exemption requires separate approval and enablement.Checked 2026-11-01. GitHub Copilot plans · GitHub credit billing
Copilot Max$100A larger published credit pool for people already using Copilot.20,000 AI credits/month including flex allotment. Credits are not tasks. Models consume them at different rates. Fable 5.1 retains prompts and outputs for safety classifiers by default; an enterprise ZDR exemption requires separate approval and enablement. Copilot Fusion has a temporary 30% discount through November 30, 2026; the recurring model-specific value excludes it. Standard Auto selection has a 10% discount in the documented Chat/cloud surfaces.Checked 2026-11-01. GitHub Copilot plans · GitHub credit billing
Google AI Plus$4.99A low-cost upgrade when free Gemini is the main limit.2x Free Gemini usage; 200 Flow credits/month Nano Banana 2.1 image generation and editing is included in the Gemini app on Google AI Plus, Pro and Ultra. Its image allowance is separate from text usage; Google does not publish a 2.1-specific image count on the access page (official US image-generation page, read October 25). Gemini uses compute-based five-hour and weekly controls; Flow credits are separate. US pricing and country eligibility apply. Gemini 4 Argon is included in AI Pro and Ultra in the Gemini app and Search. Paid Pro, eligible premium-device Pro trials and Ultra include Argon in the Antigravity app, CLI, standalone IDE and IDE extensions (official model matrix, October 27). Paid non-trial Pro and Ultra include Claude Opus 5.5 and Sonnet 5.5 thinking in Antigravity. Free and Plus do not. Gemini 3.8 Flash is available across Antigravity tiers. Older Opus/Sonnet 4.6 and GPT-OSS access there ends November 18 (official model matrix reviewed 2026-11-03). Gemini CLI and the Code Assist IDE extensions stopped serving free individual, Google AI Pro and Ultra requests on June 18, 2026. Consumer coding now uses Antigravity CLI, app and IDE extensions. Its included quota is work-dependent, with five-hour and weekly controls for paid plans; the old Gemini CLI request counts do not apply. Enterprise Code Assist and paid API-key access are separate.Checked 2026-10-31. Google AI / Gemini pricing and access · Google AI / Gemini usage documentation · Google AI / Gemini usage documentation · Google AI / Gemini usage documentation · Google AI / Gemini usage documentation
Google AI Ultra 5x$99.99Consider it when Google AI Pro runs out in the Gemini app or Antigravity.5x Pro Gemini usage; 10,000 Flow credits/month The 5x multiplier covers the Gemini app and Google Antigravity (blog.google, read 2026-09-16); Antigravity publishes no quota numbers. Jules retains its separate task allowance. Consumer Gemini CLI access ended June 18, 2026; Antigravity CLI uses the plan's shared work-dependent quota, with no published request count. Nano Banana 2.1 image generation and editing is included in the Gemini app on Google AI Plus, Pro and Ultra. Its image allowance is separate from text usage; Google does not publish a 2.1-specific image count on the access page (official US image-generation page, read October 25). Gemini uses compute-based five-hour and weekly controls; Flow credits are separate. US pricing and country eligibility apply. Gemini 4 Argon is included in AI Pro and Ultra in the Gemini app and Search. Paid Pro, eligible premium-device Pro trials and Ultra include Argon in the Antigravity app, CLI, standalone IDE and IDE extensions (official model matrix, October 27). Paid non-trial Pro and Ultra include Claude Opus 5.5 and Sonnet 5.5 thinking in Antigravity. Free and Plus do not. Gemini 3.8 Flash is available across Antigravity tiers. Older Opus/Sonnet 4.6 and GPT-OSS access there ends November 18 (official model matrix reviewed 2026-11-03). Gemini CLI and the Code Assist IDE extensions stopped serving free individual, Google AI Pro and Ultra requests on June 18, 2026. Consumer coding now uses Antigravity CLI, app and IDE extensions. Its included quota is work-dependent, with five-hour and weekly controls for paid plans; the old Gemini CLI request counts do not apply. Enterprise Code Assist and paid API-key access are separate.Checked 2026-10-31. Google AI / Gemini pricing and access · Google AI / Gemini usage documentation · Google AI / Gemini usage documentation · Google AI / Gemini usage documentation · Google AI / Gemini usage documentation
Google AI Ultra 20x$199.99Only worth it if the Gemini app or Antigravity exhausts the 5x plan; Antigravity publishes no quota numbers.20x Pro Gemini usage; 25,000 Flow credits/month The 20x multiplier covers the Gemini app and Google Antigravity (blog.google, read 2026-09-16); Antigravity publishes no quota numbers. Jules retains its separate task allowance. Consumer Gemini CLI access ended June 18, 2026; Antigravity CLI uses the plan's shared work-dependent quota, with no published request count. Nano Banana 2.1 image generation and editing is included in the Gemini app on Google AI Plus, Pro and Ultra. Its image allowance is separate from text usage; Google does not publish a 2.1-specific image count on the access page (official US image-generation page, read October 25). Gemini uses compute-based five-hour and weekly controls; Flow credits are separate. US pricing and country eligibility apply. Gemini 4 Argon is included in AI Pro and Ultra in the Gemini app and Search. Paid Pro, eligible premium-device Pro trials and Ultra include Argon in the Antigravity app, CLI, standalone IDE and IDE extensions (official model matrix, October 27). Paid non-trial Pro and Ultra include Claude Opus 5.5 and Sonnet 5.5 thinking in Antigravity. Free and Plus do not. Gemini 3.8 Flash is available across Antigravity tiers. Older Opus/Sonnet 4.6 and GPT-OSS access there ends November 18 (official model matrix reviewed 2026-11-03). Gemini CLI and the Code Assist IDE extensions stopped serving free individual, Google AI Pro and Ultra requests on June 18, 2026. Consumer coding now uses Antigravity CLI, app and IDE extensions. Its included quota is work-dependent, with five-hour and weekly controls for paid plans; the old Gemini CLI request counts do not apply. Enterprise Code Assist and paid API-key access are separate.Checked 2026-10-31. Google AI / Gemini pricing and access · Google AI / Gemini usage documentation · Google AI / Gemini usage documentation · Google AI / Gemini usage documentation · Google AI / Gemini usage documentation
Mistral Vibe Pro$14.99A lower-priced chat and coding bundle if Mistral meets your needs.Higher chat/coding limits; $25.50 monthly API credit Web/mobile chat and CLI/IDE coding are bundled, with fair-use caps. Regular Pro includes $25.50 monthly API credit; eligible education Pro has $15 instead. Team prices are per seat and the minimum purchase still needs verification. The Studio/API model catalogue offers GLM-5.3; GLM-5.2 was retired on October 31 (deprecated and retired models table, read November 1). Mistral Large 4 launched in public API preview October 6, with weights promised for later in the month. Included API usage is shared with Studio and Vibe; the preview announcement does not establish Le Chat access. Large 4 now has an exact ranked row and is included in the API access roster (official docs and launch announcement, read 2026-10-08). October 29: the API catalogue includes unmodified DeepSeek V4.1 Flash 1017 at $0.30 input, $0.007 cached input and $1.20 output per million tokens. This is an alternative use of the existing $25.50 regular Pro or $15 education API credit, not evidence of Le Chat access or an extra pool. October 30: Mistral released the improved Large 4 weights on October 29; the API model mistral-large-4 now costs $0.70 input, $0.07 cached input and $2.10 output per million tokens (launch post, API pricing table and model card, whose version label remains 26.10). Large 4 stays one alternative use of the same $25.50 regular Pro or $15 education API credit. October 31: Mistral deprecated Mistral Medium 3.5 and Mistral Large 3 (deprecation October 31, 2026; retirement April 30, 2027; replacement Mistral Large 4) and removed both from the API pricing table; they stay usable until retirement but no longer have a published rate. DeepSeek V4.1 Flash 1017 now also accepts image input at unchanged rates. The improved October 29 Mistral Large 4 (ranking row mistral-large-4-oss, High reasoning by default) is available through the Mistral API and Vibe Code, so it is one more alternative use of the same API credit, not an extra pool.Checked 2026-10-31. Mistral Vibe / Le Chat pricing and access · Mistral Vibe / Le Chat usage documentation · Mistral Vibe / Le Chat usage documentation · Mistral Vibe / Le Chat usage documentation · Mistral Vibe / Le Chat usage documentation
Perplexity Pro$20Consider it when cited search and research matter most.Search limits; no monthly Computer credits; weekly Light-effort Computer usage of unpublished size Nano Banana 2.1 generates and edits images inside Computer for subscribers with Computer access. The October 19 release note establishes image-tool access, without publishing an image allowance or model credit rate (read October 25). Ask/search and Computer use different allowances. Pro has no recurring Computer credit allocation; Max has 10,000/month. Consumer search/research caps are not fully numeric in the current comparison table. Ignore one-time credit bonuses. Paid consumer Pro plans now also include a weekly Computer usage allowance of unpublished size; without any available credits it runs tasks only on the Light effort setting, where Perplexity picks the model, and it does not roll over. Promotional Pro plans, Max and Enterprise do not get it (perplexity.ai credits help, reread 2026-10-29). The October 27 Search model table includes GPT-6.1 Sol, Gemini 3.8 Flash, Sonnet 5.5, Kimi K3, GLM-5.3, Grok 4.7 and Nemotron 3.5 Super VL; Max adds GPT-6 Astra, Gemini 4 Argon, Opus 5.5 and Fable 5.1. Search, Council, Comet and Computer use separate selectors. Computer Custom can select a coordinating model and effort when spending credits; supporting agents remain independently selected. Credits cost $1 per 100 to purchase, but model consumption rates are unpublished. One-time 4,000/35,000-credit bonuses are excluded from recurring capacity (official help, October 21). October 28 review: Max can select Ultrafast GPT-6 Astra or GPT-6.1 Sol in Computer on the web; it consumes more credits than Normal or Fast, with no published conversion. Some Free accounts receive one-time Computer turns, without a recurring refill and with one agent at a time. Current consumer prices and the October 27 Search model table remain unchanged; the credit guide was last modified October 27. The API remains separately billed. October 29: included Pro Computer usage is rolling out gradually; the account Usage page establishes availability. With credits, other effort settings including Custom are available and higher effort consumes the weekly allowance faster. The weekly quantity and model rates remain unpublished.Checked 2026-10-28. Perplexity pricing and access · Perplexity usage documentation · Perplexity usage documentation · Perplexity usage documentation · Perplexity usage documentation · Perplexity usage documentation · Perplexity usage documentation · Perplexity usage documentation
Perplexity Max$200Consider it when cited search and research matter most.10,000 Computer credits/month; higher search limits Annual subscription available through web. Nano Banana 2.1 generates and edits images inside Computer for subscribers with Computer access. The October 19 release note establishes image-tool access, without publishing an image allowance or model credit rate (read October 25). Ask/search and Computer use different allowances. Pro has no recurring Computer credit allocation; Max has 10,000/month. Consumer search/research caps are not fully numeric in the current comparison table. Ignore one-time credit bonuses. Paid consumer Pro plans now also include a weekly Computer usage allowance of unpublished size; without any available credits it runs tasks only on the Light effort setting, where Perplexity picks the model, and it does not roll over. Promotional Pro plans, Max and Enterprise do not get it (perplexity.ai credits help, reread 2026-10-29). The October 27 Search model table includes GPT-6.1 Sol, Gemini 3.8 Flash, Sonnet 5.5, Kimi K3, GLM-5.3, Grok 4.7 and Nemotron 3.5 Super VL; Max adds GPT-6 Astra, Gemini 4 Argon, Opus 5.5 and Fable 5.1. Search, Council, Comet and Computer use separate selectors. Computer Custom can select a coordinating model and effort when spending credits; supporting agents remain independently selected. Credits cost $1 per 100 to purchase, but model consumption rates are unpublished. One-time 4,000/35,000-credit bonuses are excluded from recurring capacity (official help, October 21). October 28 review: Max can select Ultrafast GPT-6 Astra or GPT-6.1 Sol in Computer on the web; it consumes more credits than Normal or Fast, with no published conversion. Some Free accounts receive one-time Computer turns, without a recurring refill and with one agent at a time. Current consumer prices and the October 27 Search model table remain unchanged; the credit guide was last modified October 27. The API remains separately billed. October 29: included Pro Computer usage is rolling out gradually; the account Usage page establishes availability. With credits, other effort settings including Custom are available and higher effort consumes the weekly allowance faster. The weekly quantity and model rates remain unpublished.Checked 2026-10-28. Perplexity pricing and access · Perplexity usage documentation · Perplexity usage documentation · Perplexity usage documentation · Perplexity usage documentation · Perplexity usage documentation · Perplexity usage documentation · Perplexity usage documentation
Kagi Ultimate$25An alternative for frequent search with research and model choice.Unlimited search; Research and flagship models; up to $25 Assistant usage/month or $270/year paid annually. Assistant accounting includes overhead: the vendor example splits $25 into $20 raw tokens and $5 operations, a stated 20% markup. Annual Assistant budget is available upfront. At exhaustion Ultimate can only renew early or wait for renewal; help no longer describes top-ups. Bundle plans such as Ultimate share a single usage limit across their products. Starter and Professional include Quick mode; Ultimate adds Research and flagship models. Assistant has a metered usage budget separate from unlimited search. Annual billing prepays the year and grants the annual Assistant budget at the start.Checked 2026-10-30. Kagi Assistant pricing and access · Kagi Assistant usage documentation · Kagi Assistant usage documentation
Abacus ChatLLM$10A low-priced way to access several model families in one app.20,000 credits/month; 3 AI Agent tasks and 3 Studio conversations (up to 2,500 credits each); chat token allowance unpublished First-month discount excluded; vendor capacity claims untested. ChatLLM Basic is $10/user/month after the introductory month, with20,000 monthly credits. Pro is $20/user/month with30,000 credits; Pro remains a catalogue candidate. Basic has three AI Agent tasks monthly; Studio has three conversations capped at2,500credits each. Pro removes those task/conversation limits, but usage still consumes credits. Official FAQ states credits are not tokens and gives only up-to token claims; purchasing10,000credits for$10 does not establish an API-dollar conversion. Advanced models and features share credits, and model fallback/attachment limits apply. Read October24.Checked 2026-10-24. Abacus ChatLLM pricing and access · Abacus ChatLLM usage documentation
Z.ai Coding Lite$18An option for coding with GLM in a supported client.2,000 credits/5h; 10,000/week Five-hour and weekly caps both apply; model and token multipliers change consumption. The developer pack documents a 50% off-peak credit rate outside Monday-Friday 14:00-18:00 Singapore time. Prepaid terms are discounted 20% quarterly and 30% yearly (browser-read 2026-09-20: Lite $18 -> $12.60, Pro $80 -> $56, Max $168 -> $117.60 per month), and require upfront commitment. A referral programme returns up to 20% in non-expiring credits. Supported coding clients may have their own separate charges.Checked 2026-11-02. Z.ai GLM Coding Plan pricing and access · Z.ai GLM Coding Plan usage documentation
Z.ai Coding Pro$80An option for coding with GLM in a supported client.12,000 credits/5h; 60,000/week Five-hour and weekly caps both apply; model and token multipliers change consumption. The developer pack documents a 50% off-peak credit rate outside Monday-Friday 14:00-18:00 Singapore time. Prepaid terms are discounted 20% quarterly and 30% yearly (browser-read 2026-09-20: Lite $18 -> $12.60, Pro $80 -> $56, Max $168 -> $117.60 per month), and require upfront commitment. A referral programme returns up to 20% in non-expiring credits. Supported coding clients may have their own separate charges.Checked 2026-11-02. Z.ai GLM Coding Plan pricing and access · Z.ai GLM Coding Plan usage documentation
Z.ai Coding Max$168An option for coding with GLM in a supported client.28,000 credits/5h; 140,000/week Five-hour and weekly caps both apply; model and token multipliers change consumption. The developer pack documents a 50% off-peak credit rate outside Monday-Friday 14:00-18:00 Singapore time. Prepaid terms are discounted 20% quarterly and 30% yearly (browser-read 2026-09-20: Lite $18 -> $12.60, Pro $80 -> $56, Max $168 -> $117.60 per month), and require upfront commitment. A referral programme returns up to 20% in non-expiring credits. Supported coding clients may have their own separate charges.Checked 2026-11-02. Z.ai GLM Coding Plan pricing and access · Z.ai GLM Coding Plan usage documentation
TRAE Pro$20A low entry price if you are comfortable working in the TRAE editor.$20 basic usage; unlimited autocomplete; up to 10 concurrent cloud tasks Monthly prices and usage values are public. Pro trial is excluded. The pricing card and the comparison table now state the same dollar figure for every paid tier, and the card's multipliers (3× Pro, 10× Pro) agree with those figures. Dollar-labelled usage follows TRAE billing rules, not cash reimbursement. October 31: TRAE's rendered model page adds GPT-6.1 Sol to the models of the current plans, at $2 input, $0.10 cache read, $2.50 cache write and $10 output per million tokens up to 272K ($4/$0.20/$5/$15 above) in both regional tables. GPT-6.1 Terra, DeepSeek V4.1 Flash and DeepSeek V4 Pro appear only in the non-U.S. table and are not added to the U.S.-dollar plan rosters. November 5: after TraeCode and TraeWork merged into TRAE, the rendered model page lists no Claude model in either regional table. The U.S.-dollar plan rosters never included Claude, and the GPT-6 Astra, GPT-6.1 Sol, GPT-6 Sol and GPT-5.6 Sol rates are unchanged (official model page, browser-read November 5).Checked 2026-10-30. TRAE pricing and access
TRAE Pro+$60A low entry price if you are comfortable working in the TRAE editor.$60 basic usage (3× Pro); up to 15 concurrent cloud tasks Monthly prices and usage values are public. Pro trial is excluded. The pricing card and the comparison table now state the same dollar figure for every paid tier, and the card's multipliers (3× Pro, 10× Pro) agree with those figures. Dollar-labelled usage follows TRAE billing rules, not cash reimbursement. October 31: TRAE's rendered model page adds GPT-6.1 Sol to the models of the current plans, at $2 input, $0.10 cache read, $2.50 cache write and $10 output per million tokens up to 272K ($4/$0.20/$5/$15 above) in both regional tables. GPT-6.1 Terra, DeepSeek V4.1 Flash and DeepSeek V4 Pro appear only in the non-U.S. table and are not added to the U.S.-dollar plan rosters. November 5: after TraeCode and TraeWork merged into TRAE, the rendered model page lists no Claude model in either regional table. The U.S.-dollar plan rosters never included Claude, and the GPT-6 Astra, GPT-6.1 Sol, GPT-6 Sol and GPT-5.6 Sol rates are unchanged (official model page, browser-read November 5).Checked 2026-10-30. TRAE pricing and access
TRAE Ultra$200A low entry price if you are comfortable working in the TRAE editor.$200 basic usage (10× Pro); model early access; up to 20 concurrent cloud tasks Monthly prices and usage values are public. Pro trial is excluded. The pricing card and the comparison table now state the same dollar figure for every paid tier, and the card's multipliers (3× Pro, 10× Pro) agree with those figures. Dollar-labelled usage follows TRAE billing rules, not cash reimbursement. October 31: TRAE's rendered model page adds GPT-6.1 Sol to the models of the current plans, at $2 input, $0.10 cache read, $2.50 cache write and $10 output per million tokens up to 272K ($4/$0.20/$5/$15 above) in both regional tables. GPT-6.1 Terra, DeepSeek V4.1 Flash and DeepSeek V4 Pro appear only in the non-U.S. table and are not added to the U.S.-dollar plan rosters. November 5: after TraeCode and TraeWork merged into TRAE, the rendered model page lists no Claude model in either regional table. The U.S.-dollar plan rosters never included Claude, and the GPT-6 Astra, GPT-6.1 Sol, GPT-6 Sol and GPT-5.6 Sol rates are unchanged (official model page, browser-read November 5).Checked 2026-10-30. TRAE pricing and access
OpenCode Go$10A modest subscription for coding in a compatible terminal client.Model-specific usage-dollar limits; 5h/weekly/monthly caps One member per workspace; compatible coding clients required. The roster includes proprietary models (Grok 4.7, GPT 5.6 Luna) beside the open ones, each with its own monthly usage-dollar cap of $15-$100, of which the 5-hour window is 20% and the weekly window 50% (checked 2026-10-17). Kimi K2.7 Code joined the roster at a $60 monthly cap by 2026-09-21; it is also available as HighSpeed on qualifying native Kimi plans as of October 14. Muse Spark 1.3 and 1.2 Contributor, each with a $60 monthly cap, are discounted in exchange for permission to train on prompts and completions and are limited to regions permitted by Meta’s Geographic Use Policy. GLM-5.1, MiniMax M2.5, Qwen3.7 Max and Qwen3.6 Plus left the roster by 2026-09-29; The current DeepSeek V4.1 Flash monthly limit is $100 (raised from $80 by October 31); the five-hour and weekly fractions still apply. Model-specific monthly usage-dollar caps have five-hour 20% and weekly 50% caps. Request-count examples assume very high cache reuse; they are not guaranteed tasks. Supported coding-client traffic requirements apply. Do not sum model limits into one spendable cash balance or assume arbitrary API automation is included. OpenCode launched Go Plus at $40/month beside Go (opencode.ai/go and opencode.ai/docs/go, read 2026-09-29): same models and token prices, with per-model monthly caps two to eight times Go's (for example Kimi K3 $15 to $60, GLM-5.3 $15 to $120, DeepSeek V4.1 Flash $80 to $160 as verified October 17); it is a catalogue candidate rather than a graded offer. Step 5 Preview is now metered at $1.00/$2.70 per million input/output tokens and $0.10 cached input, with a base monthly cap of $15 on Go and $60 on Go Plus. The 4x Step 5 Preview promotion has ended; the recurring caps remain $15/$60. Five-hour and weekly controls still apply (official Go documentation, October 16). The October 17 model table removes Grok 4.6, Kimi K2.6 and MiniMax M2.7. Beam Free is temporary and region-limited; it is not recurring included access. The October 23 model table replaces MiniMax M3.1 Flash Free with metered MiniMax M3.1 Flash: $20 monthly usage on Go and $80 on Go Plus, at $0.15 input, $0.60 output and $0.015 cached input per million tokens. Hermes clients must use version 0.21.6 or later to send session headers for Go. When included use is exhausted, pay-as-you-go requires opting into Extra usage on the Overview credits card; it is not an automatic included-plan extension (official Go documentation, October 21). The October 22 roster and token-price table were reviewed in full. The temporary LongCat 2.5 Preview route has been removed; the remaining temporary Beam Free route does not contribute a recurring model option. Peak-hour DeepSeek and the stated short-context prices are used for the measured workload; larger contexts cost more. October24: Kimi K3.1 Pro joins Go with $15/month model usage ($60 on Go Plus), at $2 input/$10 output/$0.20 cache read/$2 cache write per million. All models consume shared five-hour/weekly/monthly allowances; these are alternative uses. Global routing is enabled by default only for requests from China; other workspaces must explicitly allow Global for DeepSeek. Contributor routes require the paid-training-endpoint setting. Temporary free offers remain excluded. October 25: Gemini 4 Argon now uses the standard gemini-4-argon route, with no training use and zero-day retention. Its recurring caps remain $30 on Go and $120 on Go Plus at standard $4 input/$20 output/$0.20 cache-read prices per million tokens. The 50% discount through November 30 is temporary and excluded from recurring capacity. Muse Contributor routes retain their separate training conditions. October 30: OpenCode adds Go Teams at $20 per member per month for workspaces with several members; each member gets their own limits, $5 a month above Go on every model, and Extra usage starts switched on. It is a team catalogue candidate, not a graded offer; Go and Go Plus remain single-member workspace plans. MiniMax M3 is now priced by context: $0.30 input, $1.20 output and $0.06 cached input per million up to 512K tokens and $0.60/$2.40/$0.12 above; the recorded 152K-token request stays in the lower band. October 31: DeepSeek V4.1 Flash monthly caps rise to $100 on Go, $200 on Go Plus and $105 on Go Teams (were $80/$160/$85) at unchanged peak and off-peak prices; the table's monthly request estimates rise accordingly. No promotion or end date is stated. November 5: Hy4 joins Go beside Hy4 preview at the same $0.834 input, $2.501 output and $0.042 cached-input prices per million tokens, with $60 monthly usage on Go, $240 on Go Plus and $65 on Go Teams; Hy4 preview keeps $30/$120/$35. Hy4 is not a ranked model here and is not substituted with Hy4 preview, so it adds no token option (official Go documentation, November 5).Checked 2026-10-30. OpenCode Go pricing and access
Genspark Plus$24.99A starting point for people producing slides, documents and research in one workspace.10,000 credits/month; 50 GB storage; core-model AI Chat and AI Image free up to a limit, then credits; weekly Free Quota of unpublished size. Monthly credits expire at renewal. Agent retries and multi-model outputs consume credits. The unlimited chat and image benefit no longer applies to subscriptions started on or after September 18, 2026, so chat and image now compete with agent work for the same 10,000 credits. Plus and Pro include Nano Banana 2.1 in AI Image. AI Chat produces text only; image generation uses AI Image and its shared-credit rules. The membership page does not publish a per-image credit rate for this model (official membership benefits and AI Chat help, read October 25). Official membership page publishes monthly and annual prices. Credits reset monthly even on annual billing. From September 18, 2026 AI Chat and AI Image include core models at no credit cost only up to a limit and then draw on monthly credits, while flagship models always use credits; Plus and Pro also receive Free Quota, a weekly balance for Standard-mode agent tasks whose size is not published. Subscriptions started before September 18, 2026 keep the previous unlimited chat and image experience through December 31, 2026. Cloud Computer requires a separate subscription. The membership guide now spells out commercial use on paid plans, including client work, while requiring clear source/context acknowledgement when republishing and compliance with applicable third-party rights. This does not grant rights the service does not own (membership guide, read 2026-10-07). AI Chat also has a separate per-user five-hour fair-use cap: buying credits does not lift it, and other agents keep working while chat waits. The live model picker is the source of truth; the public model examples are not exhaustive (AI Chat help, read 2026-10-11). Buying Team keeps the existing Plus or Pro subscription running until separately canceled; Team is a separate subscription, and canceling the individual plan does not cancel Team (official membership guide, November 4).Checked 2026-10-26. Genspark pricing and access · Genspark usage documentation · Genspark usage documentation · Genspark usage documentation · Genspark usage documentation
Genspark Pro$249.99For sustained production: 12.5x the Plus credits for about 10x the monthly bill.125,000 credits/month; 1 TB storage; core-model AI Chat and AI Image free up to a higher limit, then credits; weekly Free Quota 10x the Plus balance. 12.5x the Plus credit allocation for about 10x the monthly price. Unused monthly credits expire; the higher tier pays off only if used. The unlimited chat and image benefit no longer applies to subscriptions started on or after September 18, 2026, but the 125,000-credit pool and a Free Quota ten times the Plus balance absorb that change. Plus and Pro include Nano Banana 2.1 in AI Image. AI Chat produces text only; image generation uses AI Image and its shared-credit rules. The membership page does not publish a per-image credit rate for this model (official membership benefits and AI Chat help, read October 25). Official membership page publishes monthly and annual prices. Credits reset monthly even on annual billing. From September 18, 2026 AI Chat and AI Image include core models at no credit cost only up to a limit and then draw on monthly credits, while flagship models always use credits; Plus and Pro also receive Free Quota, a weekly balance for Standard-mode agent tasks whose size is not published. Subscriptions started before September 18, 2026 keep the previous unlimited chat and image experience through December 31, 2026. Cloud Computer requires a separate subscription. The membership guide now spells out commercial use on paid plans, including client work, while requiring clear source/context acknowledgement when republishing and compliance with applicable third-party rights. This does not grant rights the service does not own (membership guide, read 2026-10-07). AI Chat also has a separate per-user five-hour fair-use cap: buying credits does not lift it, and other agents keep working while chat waits. The live model picker is the source of truth; the public model examples are not exhaustive (AI Chat help, read 2026-10-11). Buying Team keeps the existing Plus or Pro subscription running until separately canceled; Team is a separate subscription, and canceling the individual plan does not cancel Team (official membership guide, November 4).Checked 2026-10-26. Genspark pricing and access · Genspark usage documentation · Genspark usage documentation · Genspark usage documentation · Genspark usage documentation
CodeBuddy Pro$10A lower-priced editor option with explicit base and bonus credits.1,000 base + 1,000 bonus credits/month Pro includes 1,000 base plus 1,000 monthly bonus credits under the current terms. Daily activity bonuses are separate temporary promotions. Completions and automated tasks have their own limits; team credits are shared. A limited-time daily activity reward adds 30 credits a day on Free and 50 on Pro for any day the user sends a chat request; it is promotional and is not scored as recurring capacity (codebuddy.ai docs pricing, read 2026-09-20). Base and bonus credits are valid for the month they are issued and do not roll over.Checked 2026-11-02. Tencent CodeBuddy pricing and access

Full subscription catalogue

Prices retain their currency and billing basis. Annual equivalents require upfront payment. Documented means source evidence, not tested capacity or confirmed local eligibility. Unknown monthly prices, trials and retired plans are reference entries.

Service / planMonthly priceAnnual totalBuyer / price basisAllowance and conditionsEvidence
Perplexity / FreeFreeNot verifiedIndividual · standard3 Pro searches/day; 1 Research query/month Ask/search and Computer use different allowances. Pro has no recurring Computer credit allocation; Max has 10,000/month. Consumer search/research caps are not fully numeric in the current comparison table. Ignore one-time credit bonuses. Paid consumer Pro plans now also include a weekly Computer usage allowance of unpublished size; without any available credits it runs tasks only on the Light effort setting, where Perplexity picks the model, and it does not roll over. Promotional Pro plans, Max and Enterprise do not get it (perplexity.ai credits help, reread 2026-10-29). The October 27 Search model table includes GPT-6.1 Sol, Gemini 3.8 Flash, Sonnet 5.5, Kimi K3, GLM-5.3, Grok 4.7 and Nemotron 3.5 Super VL; Max adds GPT-6 Astra, Gemini 4 Argon, Opus 5.5 and Fable 5.1. Search, Council, Comet and Computer use separate selectors. Computer Custom can select a coordinating model and effort when spending credits; supporting agents remain independently selected. Credits cost $1 per 100 to purchase, but model consumption rates are unpublished. One-time 4,000/35,000-credit bonuses are excluded from recurring capacity (official help, October 21). October 28 review: Max can select Ultrafast GPT-6 Astra or GPT-6.1 Sol in Computer on the web; it consumes more credits than Normal or Fast, with no published conversion. Some Free accounts receive one-time Computer turns, without a recurring refill and with one agent at a time. Current consumer prices and the October 27 Search model table remain unchanged; the credit guide was last modified October 27. The API remains separately billed. October 29: included Pro Computer usage is rolling out gradually; the account Usage page establishes availability. With credits, other effort settings including Custom are available and higher effort consumes the weekly allowance faster. The weekly quantity and model rates remain unpublished.Documented · checked 2026-10-28 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7 · Source 8
Perplexity / ProUSD 20/monthUSD 200Individual · standardSearch limits; no monthly Computer credits; weekly Light-effort Computer usage of unpublished size Nano Banana 2.1 generates and edits images inside Computer for subscribers with Computer access. The October 19 release note establishes image-tool access, without publishing an image allowance or model credit rate (read October 25). Ask/search and Computer use different allowances. Pro has no recurring Computer credit allocation; Max has 10,000/month. Consumer search/research caps are not fully numeric in the current comparison table. Ignore one-time credit bonuses. Paid consumer Pro plans now also include a weekly Computer usage allowance of unpublished size; without any available credits it runs tasks only on the Light effort setting, where Perplexity picks the model, and it does not roll over. Promotional Pro plans, Max and Enterprise do not get it (perplexity.ai credits help, reread 2026-10-29). The October 27 Search model table includes GPT-6.1 Sol, Gemini 3.8 Flash, Sonnet 5.5, Kimi K3, GLM-5.3, Grok 4.7 and Nemotron 3.5 Super VL; Max adds GPT-6 Astra, Gemini 4 Argon, Opus 5.5 and Fable 5.1. Search, Council, Comet and Computer use separate selectors. Computer Custom can select a coordinating model and effort when spending credits; supporting agents remain independently selected. Credits cost $1 per 100 to purchase, but model consumption rates are unpublished. One-time 4,000/35,000-credit bonuses are excluded from recurring capacity (official help, October 21). October 28 review: Max can select Ultrafast GPT-6 Astra or GPT-6.1 Sol in Computer on the web; it consumes more credits than Normal or Fast, with no published conversion. Some Free accounts receive one-time Computer turns, without a recurring refill and with one agent at a time. Current consumer prices and the October 27 Search model table remain unchanged; the credit guide was last modified October 27. The API remains separately billed. October 29: included Pro Computer usage is rolling out gradually; the account Usage page establishes availability. With credits, other effort settings including Custom are available and higher effort consumes the weekly allowance faster. The weekly quantity and model rates remain unpublished.Documented · checked 2026-10-28 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7 · Source 8
Perplexity / MaxUSD 200/monthUSD 2,000Individual · standard10,000 Computer credits/month; higher search limits Annual subscription available through web. Nano Banana 2.1 generates and edits images inside Computer for subscribers with Computer access. The October 19 release note establishes image-tool access, without publishing an image allowance or model credit rate (read October 25). Ask/search and Computer use different allowances. Pro has no recurring Computer credit allocation; Max has 10,000/month. Consumer search/research caps are not fully numeric in the current comparison table. Ignore one-time credit bonuses. Paid consumer Pro plans now also include a weekly Computer usage allowance of unpublished size; without any available credits it runs tasks only on the Light effort setting, where Perplexity picks the model, and it does not roll over. Promotional Pro plans, Max and Enterprise do not get it (perplexity.ai credits help, reread 2026-10-29). The October 27 Search model table includes GPT-6.1 Sol, Gemini 3.8 Flash, Sonnet 5.5, Kimi K3, GLM-5.3, Grok 4.7 and Nemotron 3.5 Super VL; Max adds GPT-6 Astra, Gemini 4 Argon, Opus 5.5 and Fable 5.1. Search, Council, Comet and Computer use separate selectors. Computer Custom can select a coordinating model and effort when spending credits; supporting agents remain independently selected. Credits cost $1 per 100 to purchase, but model consumption rates are unpublished. One-time 4,000/35,000-credit bonuses are excluded from recurring capacity (official help, October 21). October 28 review: Max can select Ultrafast GPT-6 Astra or GPT-6.1 Sol in Computer on the web; it consumes more credits than Normal or Fast, with no published conversion. Some Free accounts receive one-time Computer turns, without a recurring refill and with one agent at a time. Current consumer prices and the October 27 Search model table remain unchanged; the credit guide was last modified October 27. The API remains separately billed. October 29: included Pro Computer usage is rolling out gradually; the account Usage page establishes availability. With credits, other effort settings including Custom are available and higher effort consumes the weekly allowance faster. The weekly quantity and model rates remain unpublished.Documented · checked 2026-10-28 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7 · Source 8
Perplexity / Education ProUSD 10/monthNot verifiedEducation · standardPro access; verify specific account allowances Student/educator verification through SheerID. Ask/search and Computer use different allowances. Pro has no recurring Computer credit allocation; Max has 10,000/month. Consumer search/research caps are not fully numeric in the current comparison table. Ignore one-time credit bonuses. Paid consumer Pro plans now also include a weekly Computer usage allowance of unpublished size; without any available credits it runs tasks only on the Light effort setting, where Perplexity picks the model, and it does not roll over. Promotional Pro plans, Max and Enterprise do not get it (perplexity.ai credits help, reread 2026-10-29). The October 27 Search model table includes GPT-6.1 Sol, Gemini 3.8 Flash, Sonnet 5.5, Kimi K3, GLM-5.3, Grok 4.7 and Nemotron 3.5 Super VL; Max adds GPT-6 Astra, Gemini 4 Argon, Opus 5.5 and Fable 5.1. Search, Council, Comet and Computer use separate selectors. Computer Custom can select a coordinating model and effort when spending credits; supporting agents remain independently selected. Credits cost $1 per 100 to purchase, but model consumption rates are unpublished. One-time 4,000/35,000-credit bonuses are excluded from recurring capacity (official help, October 21). October 28 review: Max can select Ultrafast GPT-6 Astra or GPT-6.1 Sol in Computer on the web; it consumes more credits than Normal or Fast, with no published conversion. Some Free accounts receive one-time Computer turns, without a recurring refill and with one agent at a time. Current consumer prices and the October 27 Search model table remain unchanged; the credit guide was last modified October 27. The API remains separately billed. October 29: included Pro Computer usage is rolling out gradually; the account Usage page establishes availability. With credits, other effort settings including Custom are available and higher effort consumes the weekly allowance faster. The weekly quantity and model rates remain unpublished.Conditions apply · checked 2026-10-28 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7 · Source 8
Mistral Vibe / Le Chat / FreeFreeNot verifiedIndividual · standardLimited chat, web searches and coding; basic image generation and editing; no monthly API credit advertised Web/mobile chat and CLI/IDE coding are bundled, with fair-use caps. Regular Pro includes $25.50 monthly API credit; eligible education Pro has $15 instead. Team prices are per seat and the minimum purchase still needs verification. The Studio/API model catalogue offers GLM-5.3; GLM-5.2 was retired on October 31 (deprecated and retired models table, read November 1). Mistral Large 4 launched in public API preview October 6, with weights promised for later in the month. Included API usage is shared with Studio and Vibe; the preview announcement does not establish Le Chat access. Large 4 now has an exact ranked row and is included in the API access roster (official docs and launch announcement, read 2026-10-08). October 29: the API catalogue includes unmodified DeepSeek V4.1 Flash 1017 at $0.30 input, $0.007 cached input and $1.20 output per million tokens. This is an alternative use of the existing $25.50 regular Pro or $15 education API credit, not evidence of Le Chat access or an extra pool. October 30: Mistral released the improved Large 4 weights on October 29; the API model mistral-large-4 now costs $0.70 input, $0.07 cached input and $2.10 output per million tokens (launch post, API pricing table and model card, whose version label remains 26.10). Large 4 stays one alternative use of the same $25.50 regular Pro or $15 education API credit. October 31: Mistral deprecated Mistral Medium 3.5 and Mistral Large 3 (deprecation October 31, 2026; retirement April 30, 2027; replacement Mistral Large 4) and removed both from the API pricing table; they stay usable until retirement but no longer have a published rate. DeepSeek V4.1 Flash 1017 now also accepts image input at unchanged rates. The improved October 29 Mistral Large 4 (ranking row mistral-large-4-oss, High reasoning by default) is available through the Mistral API and Vibe Code, so it is one more alternative use of the same API credit, not an extra pool.Documented · checked 2026-10-31 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
Mistral Vibe / Le Chat / ProUSD 14.99/monthNot verifiedIndividual · standardHigher chat/coding limits; $25.50 monthly API credit Web/mobile chat and CLI/IDE coding are bundled, with fair-use caps. Regular Pro includes $25.50 monthly API credit; eligible education Pro has $15 instead. Team prices are per seat and the minimum purchase still needs verification. The Studio/API model catalogue offers GLM-5.3; GLM-5.2 was retired on October 31 (deprecated and retired models table, read November 1). Mistral Large 4 launched in public API preview October 6, with weights promised for later in the month. Included API usage is shared with Studio and Vibe; the preview announcement does not establish Le Chat access. Large 4 now has an exact ranked row and is included in the API access roster (official docs and launch announcement, read 2026-10-08). October 29: the API catalogue includes unmodified DeepSeek V4.1 Flash 1017 at $0.30 input, $0.007 cached input and $1.20 output per million tokens. This is an alternative use of the existing $25.50 regular Pro or $15 education API credit, not evidence of Le Chat access or an extra pool. October 30: Mistral released the improved Large 4 weights on October 29; the API model mistral-large-4 now costs $0.70 input, $0.07 cached input and $2.10 output per million tokens (launch post, API pricing table and model card, whose version label remains 26.10). Large 4 stays one alternative use of the same $25.50 regular Pro or $15 education API credit. October 31: Mistral deprecated Mistral Medium 3.5 and Mistral Large 3 (deprecation October 31, 2026; retirement April 30, 2027; replacement Mistral Large 4) and removed both from the API pricing table; they stay usable until retirement but no longer have a published rate. DeepSeek V4.1 Flash 1017 now also accepts image input at unchanged rates. The improved October 29 Mistral Large 4 (ranking row mistral-large-4-oss, High reasoning by default) is available through the Mistral API and Vibe Code, so it is one more alternative use of the same API credit, not an extra pool.Documented · checked 2026-10-31 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
Mistral Vibe / Le Chat / Education ProUSD 5.99/monthNot verifiedEducation · standardHigher limits; $15 monthly API credit Verified eligible new students; maximum 12 months. Web/mobile chat and CLI/IDE coding are bundled, with fair-use caps. Regular Pro includes $25.50 monthly API credit; eligible education Pro has $15 instead. Team prices are per seat and the minimum purchase still needs verification. The Studio/API model catalogue offers GLM-5.3; GLM-5.2 was retired on October 31 (deprecated and retired models table, read November 1). Mistral Large 4 launched in public API preview October 6, with weights promised for later in the month. Included API usage is shared with Studio and Vibe; the preview announcement does not establish Le Chat access. Large 4 now has an exact ranked row and is included in the API access roster (official docs and launch announcement, read 2026-10-08). October 29: the API catalogue includes unmodified DeepSeek V4.1 Flash 1017 at $0.30 input, $0.007 cached input and $1.20 output per million tokens. This is an alternative use of the existing $25.50 regular Pro or $15 education API credit, not evidence of Le Chat access or an extra pool. October 30: Mistral released the improved Large 4 weights on October 29; the API model mistral-large-4 now costs $0.70 input, $0.07 cached input and $2.10 output per million tokens (launch post, API pricing table and model card, whose version label remains 26.10). Large 4 stays one alternative use of the same $25.50 regular Pro or $15 education API credit. October 31: Mistral deprecated Mistral Medium 3.5 and Mistral Large 3 (deprecation October 31, 2026; retirement April 30, 2027; replacement Mistral Large 4) and removed both from the API pricing table; they stay usable until retirement but no longer have a published rate. DeepSeek V4.1 Flash 1017 now also accepts image input at unchanged rates. The improved October 29 Mistral Large 4 (ranking row mistral-large-4-oss, High reasoning by default) is available through the Mistral API and Vibe Code, so it is one more alternative use of the same API credit, not an extra pool.Conditions apply · checked 2026-10-31 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
Mistral Vibe / Le Chat / TeamUSD 24.99/monthNot verifiedTeam / seats · per seatUp to 30GB storage/user; team controls Minimum purchase unresolved; do not treat one seat as total bill. Web/mobile chat and CLI/IDE coding are bundled, with fair-use caps. Regular Pro includes $25.50 monthly API credit; eligible education Pro has $15 instead. Team prices are per seat and the minimum purchase still needs verification. The Studio/API model catalogue offers GLM-5.3; GLM-5.2 was retired on October 31 (deprecated and retired models table, read November 1). Mistral Large 4 launched in public API preview October 6, with weights promised for later in the month. Included API usage is shared with Studio and Vibe; the preview announcement does not establish Le Chat access. Large 4 now has an exact ranked row and is included in the API access roster (official docs and launch announcement, read 2026-10-08). October 29: the API catalogue includes unmodified DeepSeek V4.1 Flash 1017 at $0.30 input, $0.007 cached input and $1.20 output per million tokens. This is an alternative use of the existing $25.50 regular Pro or $15 education API credit, not evidence of Le Chat access or an extra pool. October 30: Mistral released the improved Large 4 weights on October 29; the API model mistral-large-4 now costs $0.70 input, $0.07 cached input and $2.10 output per million tokens (launch post, API pricing table and model card, whose version label remains 26.10). Large 4 stays one alternative use of the same $25.50 regular Pro or $15 education API credit. October 31: Mistral deprecated Mistral Medium 3.5 and Mistral Large 3 (deprecation October 31, 2026; retirement April 30, 2027; replacement Mistral Large 4) and removed both from the API pricing table; they stay usable until retirement but no longer have a published rate. DeepSeek V4.1 Flash 1017 now also accepts image input at unchanged rates. The improved October 29 Mistral Large 4 (ranking row mistral-large-4-oss, High reasoning by default) is available through the Mistral API and Vibe Code, so it is one more alternative use of the same API credit, not an extra pool.Conditions apply · checked 2026-10-31 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
Grok / SuperGrok / X bundle / FreeFreeNot verifiedIndividual · standardSeparate limited Chat/Voice access SuperGrok and Plus monthly prices are published on the official comparison. Weekly capacity is shared and varies by product; absolute pool sizes are not published. Heavy pricing and annual totals remain unconfirmed. x.ai/pricing's feature comparison also lists a SuperGrok Lite tier between Free and SuperGrok, but the plan cards price only Free $0, SuperGrok $30 and SuperGrok Plus $100, so Lite is an observed unpriced tier and is not in the catalogue (read 2026-09-22). Closed to new signups on October 12, 2026. Existing subscribers retain their plan, price and features unless they choose the permanent switch to The X Subscription.Conditions apply · checked 2026-11-02 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7
Grok / SuperGrok / X bundle / SuperGrokUSD 30/monthNot verifiedIndividual · standardOne shared weekly allowance across Chat, Imagine, Voice and Build; higher limits than Free. Paid features pause at the weekly cap; free Chat/Voice remain available. Optional paid top-ups continue usage. SuperGrok and Plus monthly prices are published on the official comparison. Weekly capacity is shared and varies by product; absolute pool sizes are not published. Heavy pricing and annual totals remain unconfirmed. x.ai/pricing's feature comparison also lists a SuperGrok Lite tier between Free and SuperGrok, but the plan cards price only Free $0, SuperGrok $30 and SuperGrok Plus $100, so Lite is an observed unpriced tier and is not in the catalogue (read 2026-09-22). Closed to new signups on October 12, 2026. Existing subscribers retain their plan, price and features unless they choose the permanent switch to The X Subscription.Retired · checked 2026-10-13 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7
Grok / SuperGrok / X bundle / SuperGrok PlusUSD 100/monthNot verifiedIndividual · standardA larger shared weekly allowance across Chat, Imagine, Voice and Build. Adds priority access and 1080p video. The provider does not quantify the extra weekly capacity in its public comparison. SuperGrok and Plus monthly prices are published on the official comparison. Weekly capacity is shared and varies by product; absolute pool sizes are not published. Heavy pricing and annual totals remain unconfirmed. x.ai/pricing's feature comparison also lists a SuperGrok Lite tier between Free and SuperGrok, but the plan cards price only Free $0, SuperGrok $30 and SuperGrok Plus $100, so Lite is an observed unpriced tier and is not in the catalogue (read 2026-09-22). Closed to new signups on October 12, 2026. Existing subscribers retain their plan, price and features unless they choose the permanent switch to The X Subscription.Retired · checked 2026-10-13 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7
Grok / SuperGrok / X bundle / SuperGrok HeavyPrice not verifiedNot verifiedIndividual · standardShared weekly pool; absolute capacity not published here Current Heavy price remains unconfirmed. Published plan descriptions promise the highest limits without an absolute allowance. SuperGrok and Plus monthly prices are published on the official comparison. Weekly capacity is shared and varies by product; absolute pool sizes are not published. Heavy pricing and annual totals remain unconfirmed. x.ai/pricing's feature comparison also lists a SuperGrok Lite tier between Free and SuperGrok, but the plan cards price only Free $0, SuperGrok $30 and SuperGrok Plus $100, so Lite is an observed unpriced tier and is not in the catalogue (read 2026-09-22). Closed to new signups on October 12, 2026. Existing subscribers retain their plan, price and features unless they choose the permanent switch to The X Subscription.Retired · checked 2026-10-13 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7
Kagi Assistant / StarterUSD 5/monthUSD 54Individual · standard300 searches and 300 standard-model Assistant interactions/month; 3,600 of each/year Annual search allocation differs from monthly billing; Assistant usage is metered separately. Starter and Professional include Quick mode; Ultimate adds Research and flagship models. Assistant has a metered usage budget separate from unlimited search. Annual billing prepays the year and grants the annual Assistant budget at the start.Documented · checked 2026-11-03 Source 1 · Source 2 · Source 3
Kagi Assistant / ProfessionalUSD 10/monthUSD 108Individual · standardUnlimited search; larger Quick allowance Starter and Professional include Quick mode; Ultimate adds Research and flagship models. Assistant has a metered usage budget separate from unlimited search. Annual billing prepays the year and grants the annual Assistant budget at the start.Documented · checked 2026-11-03 Source 1 · Source 2 · Source 3
Kagi Assistant / UltimateUSD 25/monthUSD 270Individual · standardUnlimited search; Research and flagship models; up to $25 Assistant usage/month or $270/year paid annually. Assistant accounting includes overhead: the vendor example splits $25 into $20 raw tokens and $5 operations, a stated 20% markup. Annual Assistant budget is available upfront. At exhaustion Ultimate can only renew early or wait for renewal; help no longer describes top-ups. Bundle plans such as Ultimate share a single usage limit across their products. Starter and Professional include Quick mode; Ultimate adds Research and flagship models. Assistant has a metered usage budget separate from unlimited search. Annual billing prepays the year and grants the annual Assistant budget at the start.Documented · checked 2026-10-30 Source 1 · Source 2 · Source 3
Kagi Assistant / TrialFree trialNot verifiedTrial · standard100 searches and 100 Assistant interactions total Starter and Professional include Quick mode; Ultimate adds Research and flagship models. Assistant has a metered usage budget separate from unlimited search. Annual billing prepays the year and grants the annual Assistant budget at the start.Trial only · checked 2026-10-30 Source 1 · Source 2 · Source 3
Microsoft 365 / Copilot / PersonalUSD 9.99/monthUSD 99.99Individual · standardOffice/Copilot bundle; usage limits 1 TB cloud storage. Personal/Family/Premium are software-and-AI bundles, with AI benefits for the subscription owner. Family storage/seats do not create six paid AI users. Pro is monthly only. Agentic tools and premium capabilities retain usage limits.Documented · checked 2026-11-01 Source 1
Microsoft 365 / Copilot / FamilyUSD 12.99/monthUSD 129.99Individual · standardHousehold software; AI owner only 2 TB shared cloud storage; AI benefits remain owner-only. Personal/Family/Premium are software-and-AI bundles, with AI benefits for the subscription owner. Family storage/seats do not create six paid AI users. Pro is monthly only. Agentic tools and premium capabilities retain usage limits.Documented · checked 2026-11-01 Source 1
Microsoft 365 / Copilot / PremiumUSD 19.99/monthUSD 199.99Individual · standardHigher AI access; AI owner only 2 TB shared cloud storage; AI benefits remain owner-only. Personal/Family/Premium are software-and-AI bundles, with AI benefits for the subscription owner. Family storage/seats do not create six paid AI users. Pro is monthly only. Agentic tools and premium capabilities retain usage limits.Documented · checked 2026-11-01 Source 1
Microsoft 365 / Copilot / ProUSD 99.99/monthNot verifiedIndividual · standardExpanded agentic AI usage Monthly billing only. 5 TB shared cloud storage. Personal/Family/Premium are software-and-AI bundles, with AI benefits for the subscription owner. Family storage/seats do not create six paid AI users. Pro is monthly only. Agentic tools and premium capabilities retain usage limits.Documented · checked 2026-11-01 Source 1
You.com / Consumer offer unresolvedPrice not verifiedNot verifiedIndividual · standardCurrent public pricing covers APIs; a current consumer purchase price has not been established. Current /pricing and /plans expose search/research API offers. Consumer settings lead to sign-in. Older YouPro help pages do not establish a current public consumer offer. No claim that existing subscriptions have ended.Needs verification · checked 2026-09-10 Source 1 · Source 2 · Source 3
Poe / 10k/dayCAD 6.24/monthCAD 62.49Individual · standard10,000 points/day Observed CA$ label conflicts with USD footer; not eligible for USD comparison. Monthly toggle verified. Cards rendered CA$ while the same page footer says US dollars; prior reads also varied by currency. Record the observed CAD labels and hold currency-dependent comparisons. Points vary by bot, model and context length; daily points are not a monthly pool.Needs a fresh check · checked 2026-09-30 Source 1
Poe / 660k/monthCAD 24.99/monthCAD 249.99Individual · standard660,000 points/month Observed CA$ label conflicts with USD footer; not eligible for USD comparison. Monthly toggle verified. Cards rendered CA$ while the same page footer says US dollars; prior reads also varied by currency. Record the observed CAD labels and hold currency-dependent comparisons. Points vary by bot, model and context length; daily points are not a monthly pool.Needs a fresh check · checked 2026-09-30 Source 1
Poe / 1.65m/monthCAD 62.49/monthCAD 624.99Individual · standard1,650,000 points/month Observed CA$ label conflicts with USD footer; not eligible for USD comparison. Monthly toggle verified. Cards rendered CA$ while the same page footer says US dollars; prior reads also varied by currency. Record the observed CAD labels and hold currency-dependent comparisons. Points vary by bot, model and context length; daily points are not a monthly pool.Needs a fresh check · checked 2026-09-30 Source 1
Poe / 3.3m/monthCAD 124.99/monthCAD 1,249.99Individual · standard3,300,000 points/month Observed CA$ label conflicts with USD footer; not eligible for USD comparison. Monthly toggle verified. Cards rendered CA$ while the same page footer says US dollars; prior reads also varied by currency. Record the observed CAD labels and hold currency-dependent comparisons. Points vary by bot, model and context length; daily points are not a monthly pool.Needs a fresh check · checked 2026-09-30 Source 1
Poe / 8.25m/monthCAD 312.49/monthCAD 3,124.99Individual · standard8,250,000 points/month Observed CA$ label conflicts with USD footer; not eligible for USD comparison. Monthly toggle verified. Cards rendered CA$ while the same page footer says US dollars; prior reads also varied by currency. Record the observed CAD labels and hold currency-dependent comparisons. Points vary by bot, model and context length; daily points are not a monthly pool.Needs a fresh check · checked 2026-09-30 Source 1
Abacus ChatLLM / ChatLLMUSD 10/monthNot verifiedIndividual · per user20,000 credits/month; 3 AI Agent tasks and 3 Studio conversations (up to 2,500 credits each); chat token allowance unpublished First-month discount excluded; vendor capacity claims untested. ChatLLM Basic is $10/user/month after the introductory month, with20,000 monthly credits. Pro is $20/user/month with30,000 credits; Pro remains a catalogue candidate. Basic has three AI Agent tasks monthly; Studio has three conversations capped at2,500credits each. Pro removes those task/conversation limits, but usage still consumes credits. Official FAQ states credits are not tokens and gives only up-to token claims; purchasing10,000credits for$10 does not establish an API-dollar conversion. Advanced models and features share credits, and model fallback/attachment limits apply. Read October24.Documented · checked 2026-10-24 Source 1 · Source 2
Sider / NanoUSD 10/monthUSD 100Individual · standard7,500 credits/month with monthly billing; 10,000/month with yearly billing. Charged by actual usage. Temporary cloud computer available on demand. Nano, Lite, Pro and Max share access to Chat, Hand, Code and Create, with different monthly credits and cloud-computer capacity. Credits are charged by actual usage; no current model-specific token conversion is published. The older Basic/Advanced/Elite conversion tables apply only to legacy plans. All four individual tiers and both billing selectors were verified October 28 and again November 5 (rendered cards). Extra Max credits cost more than the 80,000-credit entry tier.Documented · checked 2026-11-05 Source 1 · Source 2 · Source 3
Sider / LiteUSD 25/monthUSD 250Individual · standard15,000 credits/month with monthly billing; 25,000/month with yearly billing. Charged by actual usage. Dedicated cloud computer: 2 vCPU, 2 GB RAM and 40 GB SSD. Nano, Lite, Pro and Max share access to Chat, Hand, Code and Create, with different monthly credits and cloud-computer capacity. Credits are charged by actual usage; no current model-specific token conversion is published. The older Basic/Advanced/Elite conversion tables apply only to legacy plans. All four individual tiers and both billing selectors were verified October 28 and again November 5 (rendered cards). Extra Max credits cost more than the 80,000-credit entry tier.Documented · checked 2026-11-05 Source 1 · Source 2 · Source 3
Sider / ProUSD 50/monthUSD 500Individual · standard35,000 credits/month with monthly billing; 50,000/month with yearly billing. Charged by actual usage. Dedicated cloud computer: 2 vCPU, 4 GB RAM and 40 GB SSD. Nano, Lite, Pro and Max share access to Chat, Hand, Code and Create, with different monthly credits and cloud-computer capacity. Credits are charged by actual usage; no current model-specific token conversion is published. The older Basic/Advanced/Elite conversion tables apply only to legacy plans. All four individual tiers and both billing selectors were verified October 28 and again November 5 (rendered cards). Extra Max credits cost more than the 80,000-credit entry tier.Documented · checked 2026-11-05 Source 1 · Source 2 · Source 3
Sider / MaxUSD 100/monthUSD 1,000Individual · standard80,000 credits/month with monthly billing; 100,000/month with yearly billing. Charged by actual usage. Dedicated cloud computer: 4 vCPU, 8 GB RAM and 40 GB SSD. This is the entry credit configuration; larger credit selections cost more. Nano, Lite, Pro and Max share access to Chat, Hand, Code and Create, with different monthly credits and cloud-computer capacity. Credits are charged by actual usage; no current model-specific token conversion is published. The older Basic/Advanced/Elite conversion tables apply only to legacy plans. All four individual tiers and both billing selectors were verified October 28 and again November 5 (rendered cards). Extra Max credits cost more than the 80,000-credit entry tier.Documented · checked 2026-11-05 Source 1 · Source 2 · Source 3
Monica / FreeFreeNot verifiedIndividual · standard40 basic-model accesses/day Browser verification on 2026-11-04 confirms monthly billing: Pro $9.90, Max $24.90 and Ultra $99.90. Annual prepayments are $99, $199 and $995 respectively. The Max annual flash-sale renewal price remains unconfirmed; this does not affect its separately offered monthly price. Advanced credits and personal-use restrictions still apply to unlimited chat. Paid cards include GPT-6.1 Sol, Claude Sonnet 5.5 and Gemini 4 Argon; the Opus 5.5 banner does not establish an included tier allowance.Documented · checked 2026-11-01 Source 1
Monica / ProUSD 9.90/monthUSD 99Individual · standard5,000 basic + 200 advanced queries/month; 1,500 advanced credits Browser verification on 2026-11-04 confirms monthly billing: Pro $9.90, Max $24.90 and Ultra $99.90. Annual prepayments are $99, $199 and $995 respectively. The Max annual flash-sale renewal price remains unconfirmed; this does not affect its separately offered monthly price. Advanced credits and personal-use restrictions still apply to unlimited chat. Paid cards include GPT-6.1 Sol, Claude Sonnet 5.5 and Gemini 4 Argon; the Opus 5.5 banner does not establish an included tier allowance.Conditions apply · checked 2026-11-04 Source 1
Monica / MaxUSD 24.90/monthUSD 199Individual · standardUnlimited chat claim; 4,500 advanced credits Annual flash sale $199, crossed annual price $299; annual renewal unverified. Monthly billing is separately $24.90. Browser verification on 2026-11-04 confirms monthly billing: Pro $9.90, Max $24.90 and Ultra $99.90. Annual prepayments are $99, $199 and $995 respectively. The Max annual flash-sale renewal price remains unconfirmed; this does not affect its separately offered monthly price. Advanced credits and personal-use restrictions still apply to unlimited chat. Paid cards include GPT-6.1 Sol, Claude Sonnet 5.5 and Gemini 4 Argon; the Opus 5.5 banner does not establish an included tier allowance.Conditions apply · checked 2026-10-31 Source 1
Monica / UltraUSD 99.90/monthUSD 995Individual · standardUnlimited chat claim; 20,000 advanced credits Browser verification on 2026-11-04 confirms monthly billing: Pro $9.90, Max $24.90 and Ultra $99.90. Annual prepayments are $99, $199 and $995 respectively. The Max annual flash-sale renewal price remains unconfirmed; this does not affect its separately offered monthly price. Advanced credits and personal-use restrictions still apply to unlimited chat. Paid cards include GPT-6.1 Sol, Claude Sonnet 5.5 and Gemini 4 Argon; the Opus 5.5 banner does not establish an included tier allowance.Conditions apply · checked 2026-11-04 Source 1
Z.ai GLM Coding Plan / LiteUSD 18/monthNot verifiedIndividual · standard2,000 credits/5h; 10,000/week Five-hour and weekly caps both apply; model and token multipliers change consumption. The developer pack documents a 50% off-peak credit rate outside Monday-Friday 14:00-18:00 Singapore time. Prepaid terms are discounted 20% quarterly and 30% yearly (browser-read 2026-09-20: Lite $18 -> $12.60, Pro $80 -> $56, Max $168 -> $117.60 per month), and require upfront commitment. A referral programme returns up to 20% in non-expiring credits. Supported coding clients may have their own separate charges.Documented · checked 2026-11-02 Source 1 · Source 2
Z.ai GLM Coding Plan / ProUSD 80/monthNot verifiedIndividual · standard12,000 credits/5h; 60,000/week Five-hour and weekly caps both apply; model and token multipliers change consumption. The developer pack documents a 50% off-peak credit rate outside Monday-Friday 14:00-18:00 Singapore time. Prepaid terms are discounted 20% quarterly and 30% yearly (browser-read 2026-09-20: Lite $18 -> $12.60, Pro $80 -> $56, Max $168 -> $117.60 per month), and require upfront commitment. A referral programme returns up to 20% in non-expiring credits. Supported coding clients may have their own separate charges.Documented · checked 2026-11-02 Source 1 · Source 2
Z.ai GLM Coding Plan / MaxUSD 168/monthNot verifiedIndividual · standard28,000 credits/5h; 140,000/week Five-hour and weekly caps both apply; model and token multipliers change consumption. The developer pack documents a 50% off-peak credit rate outside Monday-Friday 14:00-18:00 Singapore time. Prepaid terms are discounted 20% quarterly and 30% yearly (browser-read 2026-09-20: Lite $18 -> $12.60, Pro $80 -> $56, Max $168 -> $117.60 per month), and require upfront commitment. A referral programme returns up to 20% in non-expiring credits. Supported coding clients may have their own separate charges.Documented · checked 2026-11-02 Source 1 · Source 2
MiniMax M Plan / PlusUSD 22/monthNot verifiedIndividual · standardShared model/resource allowance; 5h + weekly caps; vendor estimate 3-4 agents Closed to new purchases. Existing auto-renewing Token Plan subscribers retain their original benefits; M Plan replaces it for new subscribers. Text, image, audio and supported video share one quota across tools. Non-video use has five-hour and weekly caps; video only the weekly cap. Absolute capacity is undisclosed, video consumes more, and unused subscription allowance does not roll over. Credit packs cost 1,000/$1 and expire after 365 days. Annual billing pays for ten months. Token Plan is closed to new purchases; existing auto-renewing subscribers retain benefits, but cannot switch old tiers or re-enable renewal after disabling it. Upgrading to M Plan is irreversible and ends historical Token Plan bonuses. M3.1 Flash replaces the preview in all three current M Plan tiers (official FAQ, October 20); this does not establish a new absolute allowance.Retired · checked 2026-10-25 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7 · Source 8
MiniMax M Plan / MaxUSD 55/monthNot verifiedIndividual · standardShared model/resource allowance; 5h + weekly caps; vendor estimate 4-5 agents Closed to new purchases. Existing auto-renewing Token Plan subscribers retain their original benefits; M Plan replaces it for new subscribers. Text, image, audio and supported video share one quota across tools. Non-video use has five-hour and weekly caps; video only the weekly cap. Absolute capacity is undisclosed, video consumes more, and unused subscription allowance does not roll over. Credit packs cost 1,000/$1 and expire after 365 days. Annual billing pays for ten months. Token Plan is closed to new purchases; existing auto-renewing subscribers retain benefits, but cannot switch old tiers or re-enable renewal after disabling it. Upgrading to M Plan is irreversible and ends historical Token Plan bonuses. M3.1 Flash replaces the preview in all three current M Plan tiers (official FAQ, October 20); this does not establish a new absolute allowance.Retired · checked 2026-10-25 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7 · Source 8
MiniMax M Plan / UltraUSD 132/monthNot verifiedIndividual · standardShared model/resource allowance; 5h + weekly caps; vendor estimate 6-7 agents Closed to new purchases. Existing auto-renewing Token Plan subscribers retain their original benefits; M Plan replaces it for new subscribers. Text, image, audio and supported video share one quota across tools. Non-video use has five-hour and weekly caps; video only the weekly cap. Absolute capacity is undisclosed, video consumes more, and unused subscription allowance does not roll over. Credit packs cost 1,000/$1 and expire after 365 days. Annual billing pays for ten months. Token Plan is closed to new purchases; existing auto-renewing subscribers retain benefits, but cannot switch old tiers or re-enable renewal after disabling it. Upgrading to M Plan is irreversible and ends historical Token Plan bonuses. M3.1 Flash replaces the preview in all three current M Plan tiers (official FAQ, October 20); this does not establish a new absolute allowance.Retired · checked 2026-10-25 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7 · Source 8
Kimi membership / Kimi Code / AndanteCNY 49/monthNot verifiedIndividual · standardAbout 30 Agent uses if the monthly pool is spent only on that feature (vendor estimate, not a guaranteed task quota); shared token-metered pool; Code rolling limits. Legacy CNY membership, retained for existing auto-renewing subscribers. New members use the new plans; migration is optional. Legacy weekly and rolling five-hour coding limits remain. The regional annual total is unverified. Standard coding uses K2.8 Preview; K3 coding requires legacy Moderato or above. Current help pages (read 2026-10-19) keep legacy benefit rules, prices and auto-renewal unchanged but no longer print per-tier Agent-use estimates; the recorded estimate comes from the September 2026 membership-pricing help. New plans and legacy subscriptions have different limits. All features share a token-metered monthly pool, with credits expiring each cycle. New Plus and above include Kimi Code and K3 with a rolling five-hour limit but no weekly limit; legacy subscriptions retain their weekly limit. K3 1M context and HighSpeed coding require new Pro or above, or legacy Allegretto or above. K3 1M coding consumes twice the quota of K3 256K. Extra Usage is separately billed. Plan eligibility follows the signed-in purchase page. Kimi Claw cloud hosts and published websites also consume the shared credits. International USD Plus/Pro/Max/Ultra pricing is published separately from the new mainland-China CNY Go/Plus/Pro/Max offers. Kimi Code has six selectable IDs: K3.1Pro and its Highspeed route, K3 and K3-256K, K2.8Preview and K2.7CodeHighspeed. An active Kimi Claw host consumes about 0.6% of monthly credits daily even while idle; a published Agent website consumes about 0.08% (official international pricing and model guide, reviewed 2026-10-14). Mainland-China Ultra is now advertised at CNY 1,399/month or CNY 13,428/year with 28× Go credits, four parallel tasks, 25 scheduled tasks/widgets, 100 projects and 50 GB storage. These regional terms do not replace the separately priced international USD Ultra offer (official regional and international pricing pages, October 20). October24 official model matrix: K3.1Pro requires newPlus+/legacyModerato+;1Mcontext requires newPro+/legacyAllegretto+. K3.1ProHighspeed requires newPro+/legacyAllegretto+, offers identical capabilities at2–3×speed and about2×quota usage. Thinking is mandatory. K3 remains selectable; the October5 measured K3 pools are retained at their original evidence date. No published conversion maps them to K3.1Pro.Retired · checked 2026-10-19 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6
Kimi membership / Kimi Code / ModeratoCNY 99/monthNot verifiedIndividual · standardAbout 60 Agent uses if the monthly pool is spent only on that feature (vendor estimate, not a guaranteed task quota); shared token-metered pool; Code rolling limits. Legacy CNY membership, retained for existing auto-renewing subscribers. New members use the new plans; migration is optional. Legacy weekly and rolling five-hour coding limits remain. The regional annual total is unverified. Current help pages (read 2026-10-19) keep legacy benefit rules, prices and auto-renewal unchanged but no longer print per-tier Agent-use estimates; the recorded estimate comes from the September 2026 membership-pricing help. New plans and legacy subscriptions have different limits. All features share a token-metered monthly pool, with credits expiring each cycle. New Plus and above include Kimi Code and K3 with a rolling five-hour limit but no weekly limit; legacy subscriptions retain their weekly limit. K3 1M context and HighSpeed coding require new Pro or above, or legacy Allegretto or above. K3 1M coding consumes twice the quota of K3 256K. Extra Usage is separately billed. Plan eligibility follows the signed-in purchase page. Kimi Claw cloud hosts and published websites also consume the shared credits. International USD Plus/Pro/Max/Ultra pricing is published separately from the new mainland-China CNY Go/Plus/Pro/Max offers. Kimi Code has six selectable IDs: K3.1Pro and its Highspeed route, K3 and K3-256K, K2.8Preview and K2.7CodeHighspeed. An active Kimi Claw host consumes about 0.6% of monthly credits daily even while idle; a published Agent website consumes about 0.08% (official international pricing and model guide, reviewed 2026-10-14). Mainland-China Ultra is now advertised at CNY 1,399/month or CNY 13,428/year with 28× Go credits, four parallel tasks, 25 scheduled tasks/widgets, 100 projects and 50 GB storage. These regional terms do not replace the separately priced international USD Ultra offer (official regional and international pricing pages, October 20). October24 official model matrix: K3.1Pro requires newPlus+/legacyModerato+;1Mcontext requires newPro+/legacyAllegretto+. K3.1ProHighspeed requires newPro+/legacyAllegretto+, offers identical capabilities at2–3×speed and about2×quota usage. Thinking is mandatory. K3 remains selectable; the October5 measured K3 pools are retained at their original evidence date. No published conversion maps them to K3.1Pro.Retired · checked 2026-10-24 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6
Kimi membership / Kimi Code / AllegrettoCNY 199/monthNot verifiedIndividual · standardAbout 150 Agent uses if the monthly pool is spent only on that feature (vendor estimate, not a guaranteed task quota); shared token-metered pool; Code rolling limits. Legacy CNY membership, retained for existing auto-renewing subscribers. New members use the new plans; migration is optional. Legacy weekly and rolling five-hour coding limits remain. The regional annual total is unverified. Current help pages (read 2026-10-19) keep legacy benefit rules, prices and auto-renewal unchanged but no longer print per-tier Agent-use estimates; the recorded estimate comes from the September 2026 membership-pricing help. New plans and legacy subscriptions have different limits. All features share a token-metered monthly pool, with credits expiring each cycle. New Plus and above include Kimi Code and K3 with a rolling five-hour limit but no weekly limit; legacy subscriptions retain their weekly limit. K3 1M context and HighSpeed coding require new Pro or above, or legacy Allegretto or above. K3 1M coding consumes twice the quota of K3 256K. Extra Usage is separately billed. Plan eligibility follows the signed-in purchase page. Kimi Claw cloud hosts and published websites also consume the shared credits. International USD Plus/Pro/Max/Ultra pricing is published separately from the new mainland-China CNY Go/Plus/Pro/Max offers. Kimi Code has six selectable IDs: K3.1Pro and its Highspeed route, K3 and K3-256K, K2.8Preview and K2.7CodeHighspeed. An active Kimi Claw host consumes about 0.6% of monthly credits daily even while idle; a published Agent website consumes about 0.08% (official international pricing and model guide, reviewed 2026-10-14). Mainland-China Ultra is now advertised at CNY 1,399/month or CNY 13,428/year with 28× Go credits, four parallel tasks, 25 scheduled tasks/widgets, 100 projects and 50 GB storage. These regional terms do not replace the separately priced international USD Ultra offer (official regional and international pricing pages, October 20). October24 official model matrix: K3.1Pro requires newPlus+/legacyModerato+;1Mcontext requires newPro+/legacyAllegretto+. K3.1ProHighspeed requires newPro+/legacyAllegretto+, offers identical capabilities at2–3×speed and about2×quota usage. Thinking is mandatory. K3 remains selectable; the October5 measured K3 pools are retained at their original evidence date. No published conversion maps them to K3.1Pro.Retired · checked 2026-10-24 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6
Kimi membership / Kimi Code / AllegroCNY 699/monthNot verifiedIndividual · standardAbout 360 Agent uses if the monthly pool is spent only on that feature (vendor estimate, not a guaranteed task quota); shared token-metered pool; Code rolling limits. Legacy CNY membership, retained for existing auto-renewing subscribers. New members use the new plans; migration is optional. Legacy weekly and rolling five-hour coding limits remain. The regional annual total is unverified. Current help pages (read 2026-10-19) keep legacy benefit rules, prices and auto-renewal unchanged but no longer print per-tier Agent-use estimates; the recorded estimate comes from the September 2026 membership-pricing help. New plans and legacy subscriptions have different limits. All features share a token-metered monthly pool, with credits expiring each cycle. New Plus and above include Kimi Code and K3 with a rolling five-hour limit but no weekly limit; legacy subscriptions retain their weekly limit. K3 1M context and HighSpeed coding require new Pro or above, or legacy Allegretto or above. K3 1M coding consumes twice the quota of K3 256K. Extra Usage is separately billed. Plan eligibility follows the signed-in purchase page. Kimi Claw cloud hosts and published websites also consume the shared credits. International USD Plus/Pro/Max/Ultra pricing is published separately from the new mainland-China CNY Go/Plus/Pro/Max offers. Kimi Code has six selectable IDs: K3.1Pro and its Highspeed route, K3 and K3-256K, K2.8Preview and K2.7CodeHighspeed. An active Kimi Claw host consumes about 0.6% of monthly credits daily even while idle; a published Agent website consumes about 0.08% (official international pricing and model guide, reviewed 2026-10-14). Mainland-China Ultra is now advertised at CNY 1,399/month or CNY 13,428/year with 28× Go credits, four parallel tasks, 25 scheduled tasks/widgets, 100 projects and 50 GB storage. These regional terms do not replace the separately priced international USD Ultra offer (official regional and international pricing pages, October 20). October24 official model matrix: K3.1Pro requires newPlus+/legacyModerato+;1Mcontext requires newPro+/legacyAllegretto+. K3.1ProHighspeed requires newPro+/legacyAllegretto+, offers identical capabilities at2–3×speed and about2×quota usage. Thinking is mandatory. K3 remains selectable; the October5 measured K3 pools are retained at their original evidence date. No published conversion maps them to K3.1Pro.Retired · checked 2026-10-24 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6
Alibaba Cloud Coding Plan / ProUSD 50/monthNot verifiedIndividual · standard6,000 calls/5h; 45,000/week; 90,000/month Limited purchase slots; plan-specific API key and endpoint. Pro purchase slots are limited, restocked at midnight UTC+8. Only explicitly listed model versions qualify. Model calls, not completed coding tasks, consume quota. All three limits apply simultaneously. Lite cannot be newly subscribed or renewed. November 5: Alibaba Cloud documentation now files this page as "Coding Plan (legacy)" beside a new credit-billed Token Plan for personal and team use with monthly, quarterly or yearly billing. The Pro page still offers $50 a month with the same limits and limited daily slots (official documentation, November 5).Conditions apply · checked 2026-11-02 Source 1
Alibaba Cloud Coding Plan / LitePrice not verifiedNot verifiedIndividual · standardNew subscriptions stopped March 20; renewals stopped April 13, 2026. Pro purchase slots are limited, restocked at midnight UTC+8. Only explicitly listed model versions qualify. Model calls, not completed coding tasks, consume quota. All three limits apply simultaneously. Lite cannot be newly subscribed or renewed. November 5: Alibaba Cloud documentation now files this page as "Coding Plan (legacy)" beside a new credit-billed Token Plan for personal and team use with monthly, quarterly or yearly billing. The Pro page still offers $50 a month with the same limits and limited daily slots (official documentation, November 5).Retired · checked 2026-10-25 Source 1
Devin / Windsurf / FreeFreeNot verifiedIndividual · standardLight agent allowance; limited model selection Windsurf pricing redirects to Devin. Current individual offers have daily/weekly allowances without numeric capacity. Extra usage is sold at API rates. Team pricing is $80 base plus $40/full developer seat; brand history must not create duplicate offers.Documented · checked 2026-11-01 Source 1 · Source 2
Devin / Windsurf / ProUSD 20/monthNot verifiedIndividual · standardDaily/weekly pool; frontier and cloud agents Windsurf pricing redirects to Devin. Current individual offers have daily/weekly allowances without numeric capacity. Extra usage is sold at API rates. Team pricing is $80 base plus $40/full developer seat; brand history must not create duplicate offers.Documented · checked 2026-11-01 Source 1 · Source 2
Devin / Windsurf / MaxUSD 200/monthNot verifiedIndividual · standardHigher daily/weekly pool; numeric size unspecified Windsurf pricing redirects to Devin. Current individual offers have daily/weekly allowances without numeric capacity. Extra usage is sold at API rates. Team pricing is $80 base plus $40/full developer seat; brand history must not create duplicate offers.Documented · checked 2026-11-01 Source 1 · Source 2
Kiro / FreeFreeNot verifiedIndividual · standard50 credits/month Credits are weighted by task and model. IDE/CLI/web usage shares the balance, and web compute has no separate charge. $0.04 add-on credit pricing. The first-upgrade free trial is excluded: from November 1 the rest of the first month is free plus 500 trial credits usable within 12 months, replacing the $20 sign-up credit (pricing FAQ, read 2026-11-01). Country and provider model restrictions apply. Subscriptions cover Kiro IDE, CLI, web, Crew, ACP-compatible IDEs and CI/CD automation; use through third-party automation harnesses that route requests outside Kiro's native interfaces is not permitted (pricing FAQ, read 2026-09-16). Kimi K3 is included on paid plans at a 2.4x credit multiplier with a 1M context window; individual subscriptions route it through the US endpoint (official model matrix, October 16). DeepSeek 3.2 and GLM-5 use the US endpoint only; MiniMax M2.5 and M2.1 support both US and EU endpoints (official model matrix, October 20). No absolute token conversion follows from the unchanged weighted-credit counts. October24: Fable5.1 is generally available only for KiroEnterprise under administrator model governance, with6×credits and US-East-1 inference. This does not extend personal-plan access. October 25: Claude Haiku 5.5 is available on Pro, Pro+, Pro Max and Power, with a 1M context window and a 0.06x credit multiplier through 100K tokens, rising to 0.3x above 100K. US and EU endpoints are listed; individual subscribers use the US endpoint. Credits still depend on task size, output and thinking, so these relative rates do not establish tokens per credit. November 5: teams of two or more now buy credit pooling at $20 per seat per month; each seat adds 500 included credits to a shared pool, add-on credits cost $0.04 and last 12 months, and unused included credits expire monthly. It is a team catalogue candidate; the individual Free, Pro, Pro+, Pro Max and Power tiers remain available unchanged (official pricing page, November 5).Documented · checked 2026-11-01 Source 1 · Source 2
Kiro / ProUSD 20/monthNot verifiedIndividual · standard1,000 credits/month Credits are weighted by task and model. IDE/CLI/web usage shares the balance, and web compute has no separate charge. $0.04 add-on credit pricing. The first-upgrade free trial is excluded: from November 1 the rest of the first month is free plus 500 trial credits usable within 12 months, replacing the $20 sign-up credit (pricing FAQ, read 2026-11-01). Country and provider model restrictions apply. Subscriptions cover Kiro IDE, CLI, web, Crew, ACP-compatible IDEs and CI/CD automation; use through third-party automation harnesses that route requests outside Kiro's native interfaces is not permitted (pricing FAQ, read 2026-09-16). Kimi K3 is included on paid plans at a 2.4x credit multiplier with a 1M context window; individual subscriptions route it through the US endpoint (official model matrix, October 16). DeepSeek 3.2 and GLM-5 use the US endpoint only; MiniMax M2.5 and M2.1 support both US and EU endpoints (official model matrix, October 20). No absolute token conversion follows from the unchanged weighted-credit counts. October24: Fable5.1 is generally available only for KiroEnterprise under administrator model governance, with6×credits and US-East-1 inference. This does not extend personal-plan access. October 25: Claude Haiku 5.5 is available on Pro, Pro+, Pro Max and Power, with a 1M context window and a 0.06x credit multiplier through 100K tokens, rising to 0.3x above 100K. US and EU endpoints are listed; individual subscribers use the US endpoint. Credits still depend on task size, output and thinking, so these relative rates do not establish tokens per credit. November 5: teams of two or more now buy credit pooling at $20 per seat per month; each seat adds 500 included credits to a shared pool, add-on credits cost $0.04 and last 12 months, and unused included credits expire monthly. It is a team catalogue candidate; the individual Free, Pro, Pro+, Pro Max and Power tiers remain available unchanged (official pricing page, November 5).Documented · checked 2026-11-01 Source 1 · Source 2
Kiro / Pro+USD 40/monthNot verifiedIndividual · standard2,000 credits/month Credits are weighted by task and model. IDE/CLI/web usage shares the balance, and web compute has no separate charge. $0.04 add-on credit pricing. The first-upgrade free trial is excluded: from November 1 the rest of the first month is free plus 500 trial credits usable within 12 months, replacing the $20 sign-up credit (pricing FAQ, read 2026-11-01). Country and provider model restrictions apply. Subscriptions cover Kiro IDE, CLI, web, Crew, ACP-compatible IDEs and CI/CD automation; use through third-party automation harnesses that route requests outside Kiro's native interfaces is not permitted (pricing FAQ, read 2026-09-16). Kimi K3 is included on paid plans at a 2.4x credit multiplier with a 1M context window; individual subscriptions route it through the US endpoint (official model matrix, October 16). DeepSeek 3.2 and GLM-5 use the US endpoint only; MiniMax M2.5 and M2.1 support both US and EU endpoints (official model matrix, October 20). No absolute token conversion follows from the unchanged weighted-credit counts. October24: Fable5.1 is generally available only for KiroEnterprise under administrator model governance, with6×credits and US-East-1 inference. This does not extend personal-plan access. October 25: Claude Haiku 5.5 is available on Pro, Pro+, Pro Max and Power, with a 1M context window and a 0.06x credit multiplier through 100K tokens, rising to 0.3x above 100K. US and EU endpoints are listed; individual subscribers use the US endpoint. Credits still depend on task size, output and thinking, so these relative rates do not establish tokens per credit. November 5: teams of two or more now buy credit pooling at $20 per seat per month; each seat adds 500 included credits to a shared pool, add-on credits cost $0.04 and last 12 months, and unused included credits expire monthly. It is a team catalogue candidate; the individual Free, Pro, Pro+, Pro Max and Power tiers remain available unchanged (official pricing page, November 5).Documented · checked 2026-11-01 Source 1 · Source 2
Kiro / Pro MaxUSD 100/monthNot verifiedIndividual · standard5,000 credits/month Credits are weighted by task and model. IDE/CLI/web usage shares the balance, and web compute has no separate charge. $0.04 add-on credit pricing. The first-upgrade free trial is excluded: from November 1 the rest of the first month is free plus 500 trial credits usable within 12 months, replacing the $20 sign-up credit (pricing FAQ, read 2026-11-01). Country and provider model restrictions apply. Subscriptions cover Kiro IDE, CLI, web, Crew, ACP-compatible IDEs and CI/CD automation; use through third-party automation harnesses that route requests outside Kiro's native interfaces is not permitted (pricing FAQ, read 2026-09-16). Kimi K3 is included on paid plans at a 2.4x credit multiplier with a 1M context window; individual subscriptions route it through the US endpoint (official model matrix, October 16). DeepSeek 3.2 and GLM-5 use the US endpoint only; MiniMax M2.5 and M2.1 support both US and EU endpoints (official model matrix, October 20). No absolute token conversion follows from the unchanged weighted-credit counts. October24: Fable5.1 is generally available only for KiroEnterprise under administrator model governance, with6×credits and US-East-1 inference. This does not extend personal-plan access. October 25: Claude Haiku 5.5 is available on Pro, Pro+, Pro Max and Power, with a 1M context window and a 0.06x credit multiplier through 100K tokens, rising to 0.3x above 100K. US and EU endpoints are listed; individual subscribers use the US endpoint. Credits still depend on task size, output and thinking, so these relative rates do not establish tokens per credit. November 5: teams of two or more now buy credit pooling at $20 per seat per month; each seat adds 500 included credits to a shared pool, add-on credits cost $0.04 and last 12 months, and unused included credits expire monthly. It is a team catalogue candidate; the individual Free, Pro, Pro+, Pro Max and Power tiers remain available unchanged (official pricing page, November 5).Documented · checked 2026-11-01 Source 1 · Source 2
Kiro / PowerUSD 200/monthNot verifiedIndividual · standard10,000 credits/month Credits are weighted by task and model. IDE/CLI/web usage shares the balance, and web compute has no separate charge. $0.04 add-on credit pricing. The first-upgrade free trial is excluded: from November 1 the rest of the first month is free plus 500 trial credits usable within 12 months, replacing the $20 sign-up credit (pricing FAQ, read 2026-11-01). Country and provider model restrictions apply. Subscriptions cover Kiro IDE, CLI, web, Crew, ACP-compatible IDEs and CI/CD automation; use through third-party automation harnesses that route requests outside Kiro's native interfaces is not permitted (pricing FAQ, read 2026-09-16). Kimi K3 is included on paid plans at a 2.4x credit multiplier with a 1M context window; individual subscriptions route it through the US endpoint (official model matrix, October 16). DeepSeek 3.2 and GLM-5 use the US endpoint only; MiniMax M2.5 and M2.1 support both US and EU endpoints (official model matrix, October 20). No absolute token conversion follows from the unchanged weighted-credit counts. October24: Fable5.1 is generally available only for KiroEnterprise under administrator model governance, with6×credits and US-East-1 inference. This does not extend personal-plan access. October 25: Claude Haiku 5.5 is available on Pro, Pro+, Pro Max and Power, with a 1M context window and a 0.06x credit multiplier through 100K tokens, rising to 0.3x above 100K. US and EU endpoints are listed; individual subscribers use the US endpoint. Credits still depend on task size, output and thinking, so these relative rates do not establish tokens per credit. November 5: teams of two or more now buy credit pooling at $20 per seat per month; each seat adds 500 included credits to a shared pool, add-on credits cost $0.04 and last 12 months, and unused included credits expire monthly. It is a team catalogue candidate; the individual Free, Pro, Pro+, Pro Max and Power tiers remain available unchanged (official pricing page, November 5).Documented · checked 2026-11-01 Source 1 · Source 2
Zed / PersonalFreeNot verifiedIndividual · standardExternal agents and bring-your-own provider keys; no Zed-hosted Edit Predictions Hosted Edit Predictions require Zed Pro; Free can use supported predictions from other providers and external agents. Provider charges and allowances apply separately. Pro bundles only $5 hosted-model usage, then bills at API list price +10%. Personal can use external agents/BYOK whose bills are separate. Business buys administration and includes no fixed hosted credit pool; it is not a capacity upgrade over Pro.Documented · checked 2026-11-01 Source 1 · Source 2
Zed / ProUSD 10/monthNot verifiedIndividual · standard$5 monthly hosted tokens; unlimited edit predictions Pro bundles only $5 hosted-model usage, then bills at API list price +10%. Personal can use external agents/BYOK whose bills are separate. Business buys administration and includes no fixed hosted credit pool; it is not a capacity upgrade over Pro.Documented · checked 2026-11-01 Source 1 · Source 2
Zed / BusinessUSD 30/monthNot verifiedTeam / seats · per seatNo fixed LLM credits; governance + edit predictions Pro bundles only $5 hosted-model usage, then bills at API list price +10%. Personal can use external agents/BYOK whose bills are separate. Business buys administration and includes no fixed hosted credit pool; it is not a capacity upgrade over Pro.Documented · checked 2026-11-01 Source 1 · Source 2
JetBrains AI / Junie / AI FreeFreeNot verifiedIndividual · standard3 AI credits/30 days Check IDE license and existing bundled entitlement; one AI credit references $1. Individual prices/credits differ from organizational offers. IDE/product eligibility matters: AI Free is unavailable in several free editions and Android Studio; existing All Products Pack/dotUltimate/DataSpell may already include Pro. Region and license-server restrictions apply.Conditions apply · checked 2026-10-30 Source 1
JetBrains AI / Junie / AI ProUSD 10/monthNot verifiedIndividual · standard10 AI credits/30 days Check IDE license and existing bundled entitlement; one AI credit references $1. Individual prices/credits differ from organizational offers. IDE/product eligibility matters: AI Free is unavailable in several free editions and Android Studio; existing All Products Pack/dotUltimate/DataSpell may already include Pro. Region and license-server restrictions apply.Conditions apply · checked 2026-10-30 Source 1
JetBrains AI / Junie / AI UltimateUSD 30/monthNot verifiedIndividual · standard35 AI credits/30 days Check IDE license and existing bundled entitlement; one AI credit references $1. Individual prices/credits differ from organizational offers. IDE/product eligibility matters: AI Free is unavailable in several free editions and Android Studio; existing All Products Pack/dotUltimate/DataSpell may already include Pro. Region and license-server restrictions apply.Conditions apply · checked 2026-10-30 Source 1
Amp / MegawattUSD 20/monthNot verifiedIndividual · standardIncluded Amp inference credits (amount unspecified) plus 45,000 minutes of orbin’ time A paid tier includes agent usage (model inference) and orb usage plus every Amp product feature; the pricing card no longer prints the included credit amount, and the docs pricing page states the inclusion instead (both read 2026-09-16). Included usage resets each billing period and does not roll over. Bringing your own key or another subscription carries no Amp token fee, but that provider bills you separately. Pay-as-you-go mode remains an alternative; extra monthly usage is not guaranteed. Paid tiers now advertise unlimited public/private repositories and Teams at no extra charge. The October24 Amp Credits matrix assigns Low to GLM5.3Flash High, Medium to Opus5.5 High, High to Astra Medium and Ultra to Fable5.1 High. Puck uses GPT5.6Sol; Task and Read Thread subagents use the thread model. Linked subscriptions bill their own allowances. The expanded pricing FAQ confirms both tiers include Amp inference credits, but current pricing documentation does not quantify the included credit amount (browser-verified 2026-10-08). October 29: both rendered tier selectors and the expanded pricing FAQ still confirm $20/$200, 45,000/480,000 orb minutes and included inference credits. The credit amount remains unspecified; linked subscriptions or API keys use their own paid allowances.Documented · checked 2026-11-05 Source 1 · Source 2 · Source 3
Amp / GigawattUSD 200/monthNot verifiedIndividual · standardIncluded Amp inference credits (amount unspecified) plus 480,000 minutes of orbin’ time A paid tier includes agent usage (model inference) and orb usage plus every Amp product feature; the pricing card no longer prints the included credit amount, and the docs pricing page states the inclusion instead (both read 2026-09-16). Included usage resets each billing period and does not roll over. Bringing your own key or another subscription carries no Amp token fee, but that provider bills you separately. Pay-as-you-go mode remains an alternative; extra monthly usage is not guaranteed. Paid tiers now advertise unlimited public/private repositories and Teams at no extra charge. The October24 Amp Credits matrix assigns Low to GLM5.3Flash High, Medium to Opus5.5 High, High to Astra Medium and Ultra to Fable5.1 High. Puck uses GPT5.6Sol; Task and Read Thread subagents use the thread model. Linked subscriptions bill their own allowances. The expanded pricing FAQ confirms both tiers include Amp inference credits, but current pricing documentation does not quantify the included credit amount (browser-verified 2026-10-08). October 29: both rendered tier selectors and the expanded pricing FAQ still confirm $20/$200, 45,000/480,000 orb minutes and included inference credits. The credit amount remains unspecified; linked subscriptions or API keys use their own paid allowances.Documented · checked 2026-11-05 Source 1 · Source 2 · Source 3
Factory / ProUSD 20/monthNot verifiedIndividual · standardBase rolling allowance Individual tiers publish relative rolling allowances, not guaranteed token/task totals. Teams cost $60 base plus $40/seat with up to ten seats; do not mislabel the seat price as the total. Included cloud-computer access differs across tiers. Fable 5.1 and Fable 5 use a 4x multiplier and require administrative opt-in to their 30-day data retention terms. Model access and these conditions are documented in the native model catalogue (October 15). Mistral Large 4 now costs 0.544× after its 0.272× promotion ended. New speed routes include GPT-6.1 Sol Ultra Fast 4.8×, GPT-6 Sol Fast 1.6× and GPT-6 Luna Fast 0.08× (normal 0.04×). The catalogue also lists Nemotron 3.5 Super 0.08×, Muse Spark 1.2 0.5× and Beam 1.0 0.52×. These multipliers do not reveal the underlying rolling pool (official model catalogue, October 21). The October 22 native model catalogue removes Muse Spark Open; Spark 1.2 remains included. The rolling allowance is still unpublished. October24 model matrix adds DeepSeekV4.1Flash1017 (0.12×, off/low/high/max) and MiniMaxM3.1Flash (0.06×, low/medium/high/xhigh/max). These multipliers do not quantify the base rolling allowance. October 28: Muse Spark 1.3 is included at a 0.5× usage multiplier, with low/medium/high/xhigh effort. The underlying rolling token allowance remains unpublished (native model catalogue).Documented · checked 2026-10-31 Source 1 · Source 2
Factory / PlusUSD 100/monthNot verifiedIndividual · standardApproximately 5x Pro usage Individual tiers publish relative rolling allowances, not guaranteed token/task totals. Teams cost $60 base plus $40/seat with up to ten seats; do not mislabel the seat price as the total. Included cloud-computer access differs across tiers. Fable 5.1 and Fable 5 use a 4x multiplier and require administrative opt-in to their 30-day data retention terms. Model access and these conditions are documented in the native model catalogue (October 15). Mistral Large 4 now costs 0.544× after its 0.272× promotion ended. New speed routes include GPT-6.1 Sol Ultra Fast 4.8×, GPT-6 Sol Fast 1.6× and GPT-6 Luna Fast 0.08× (normal 0.04×). The catalogue also lists Nemotron 3.5 Super 0.08×, Muse Spark 1.2 0.5× and Beam 1.0 0.52×. These multipliers do not reveal the underlying rolling pool (official model catalogue, October 21). The October 22 native model catalogue removes Muse Spark Open; Spark 1.2 remains included. The rolling allowance is still unpublished. October24 model matrix adds DeepSeekV4.1Flash1017 (0.12×, off/low/high/max) and MiniMaxM3.1Flash (0.06×, low/medium/high/xhigh/max). These multipliers do not quantify the base rolling allowance. October 28: Muse Spark 1.3 is included at a 0.5× usage multiplier, with low/medium/high/xhigh effort. The underlying rolling token allowance remains unpublished (native model catalogue).Documented · checked 2026-10-31 Source 1 · Source 2
Factory / MaxUSD 200/monthNot verifiedIndividual · standardApproximately 10x Pro usage Individual tiers publish relative rolling allowances, not guaranteed token/task totals. Teams cost $60 base plus $40/seat with up to ten seats; do not mislabel the seat price as the total. Included cloud-computer access differs across tiers. Fable 5.1 and Fable 5 use a 4x multiplier and require administrative opt-in to their 30-day data retention terms. Model access and these conditions are documented in the native model catalogue (October 15). Mistral Large 4 now costs 0.544× after its 0.272× promotion ended. New speed routes include GPT-6.1 Sol Ultra Fast 4.8×, GPT-6 Sol Fast 1.6× and GPT-6 Luna Fast 0.08× (normal 0.04×). The catalogue also lists Nemotron 3.5 Super 0.08×, Muse Spark 1.2 0.5× and Beam 1.0 0.52×. These multipliers do not reveal the underlying rolling pool (official model catalogue, October 21). The October 22 native model catalogue removes Muse Spark Open; Spark 1.2 remains included. The rolling allowance is still unpublished. October24 model matrix adds DeepSeekV4.1Flash1017 (0.12×, off/low/high/max) and MiniMaxM3.1Flash (0.06×, low/medium/high/xhigh/max). These multipliers do not quantify the base rolling allowance. October 28: Muse Spark 1.3 is included at a 0.5× usage multiplier, with low/medium/high/xhigh effort. The underlying rolling token allowance remains unpublished (native model catalogue).Documented · checked 2026-10-31 Source 1 · Source 2
Replit / StarterFreeNot verifiedIndividual · standardDaily Agent and monthly Cloud grants; quotas not fully numeric Core and entry Pro monthly prices are $20 and $100; annual equivalents are $18 and $90. Credits also pay hosting, storage, transfer and integrations. Pro pools usage across up to 15 builders; the Pro docs now state that unused credits roll over for two months (docs.replit.com/billing/plans/replit-pro, read 2026-10-02), where the March launch blog said one month; the launch blog's Turbo mode ("2x faster than Power ... up to 6x more") belonged to the retired Economy/Power/Turbo mode set; the current docs offer Free, Power and Max modes and reserve Fast model variants (such as Claude Opus 5 Fast and GPT-6 Luna Fast) for Pro and Enterprise without publishing their credit premium (docs.replit.com agent-modes and model-selector, read 2026-09-28). Higher credit tiers are separate configurations.Documented · checked 2026-11-01 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
Replit / CoreUSD 20/monthUSD 216Individual · standard$20 monthly paid-service credit + Free Mode allowance Monthly price confirmed by the official plan announcement, consistent with the current pricing card. Annual total is $18 x 12. Includes up to 5 builders; paid-service credits also cover hosting and integrations. At most 1 active background Agent task concurrently (plan documentation, 2026-10-11). Core and entry Pro monthly prices are $20 and $100; annual equivalents are $18 and $90. Credits also pay hosting, storage, transfer and integrations. Pro pools usage across up to 15 builders; the Pro docs now state that unused credits roll over for two months (docs.replit.com/billing/plans/replit-pro, read 2026-10-02), where the March launch blog said one month; the launch blog's Turbo mode ("2x faster than Power ... up to 6x more") belonged to the retired Economy/Power/Turbo mode set; the current docs offer Free, Power and Max modes and reserve Fast model variants (such as Claude Opus 5 Fast and GPT-6 Luna Fast) for Pro and Enterprise without publishing their credit premium (docs.replit.com agent-modes and model-selector, read 2026-09-28). Higher credit tiers are separate configurations.Documented · checked 2026-11-02 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
Replit / Pro entry tierUSD 100/monthUSD 1,080Individual · standard$100 monthly paid-service credit + higher Free Mode allowance Entry tier: $100 monthly or $90 x 12 annually; up to 15 builders share credits without per-user fees. Unused credits roll over for two months under the current Pro docs (the launch blog said one month). Pro-only Fast model variants cost more credits at an unpublished premium; higher tiers cost extra. At most 10 active background Agent tasks concurrently (plan documentation, 2026-10-11). Core and entry Pro monthly prices are $20 and $100; annual equivalents are $18 and $90. Credits also pay hosting, storage, transfer and integrations. Pro pools usage across up to 15 builders; the Pro docs now state that unused credits roll over for two months (docs.replit.com/billing/plans/replit-pro, read 2026-10-02), where the March launch blog said one month; the launch blog's Turbo mode ("2x faster than Power ... up to 6x more") belonged to the retired Economy/Power/Turbo mode set; the current docs offer Free, Power and Max modes and reserve Fast model variants (such as Claude Opus 5 Fast and GPT-6 Luna Fast) for Pro and Enterprise without publishing their credit premium (docs.replit.com agent-modes and model-selector, read 2026-09-28). Higher credit tiers are separate configurations.Documented · checked 2026-11-02 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
Lovable / FreeFreeNot verifiedWorkspace · standard5 build credits/day, capped 30/month; Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-10-31 Source 1 · Source 2 · Source 3
Lovable / Pro 100 creditsUSD 25/monthUSD 250Workspace · standard100 Pro credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Pro 200 creditsUSD 50/monthUSD 500Workspace · standard200 Pro credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Pro 400 creditsUSD 100/monthUSD 1,000Workspace · standard400 Pro credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Pro 800 creditsUSD 200/monthUSD 2,000Workspace · standard800 Pro credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Pro 1200 creditsUSD 294/monthUSD 2,940Workspace · standard1,200 Pro credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Pro 2000 creditsUSD 480/monthUSD 4,800Workspace · standard2,000 Pro credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Pro 3000 creditsUSD 705/monthUSD 7,050Workspace · standard3,000 Pro credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Pro 4000 creditsUSD 920/monthUSD 9,200Workspace · standard4,000 Pro credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Pro 5000 creditsUSD 1,125/monthUSD 11,250Workspace · standard5,000 Pro credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Pro 7500 creditsUSD 1,688/monthUSD 16,880Workspace · standard7,500 Pro credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Pro 10000 creditsUSD 2,250/monthUSD 22,500Workspace · standard10,000 Pro credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Business 100 creditsUSD 50/monthUSD 500Workspace · standard100 Business credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Business 200 creditsUSD 100/monthUSD 1,000Workspace · standard200 Business credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Business 400 creditsUSD 200/monthUSD 2,000Workspace · standard400 Business credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Business 800 creditsUSD 400/monthUSD 4,000Workspace · standard800 Business credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Business 1200 creditsUSD 588/monthUSD 5,880Workspace · standard1,200 Business credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Business 2000 creditsUSD 960/monthUSD 9,600Workspace · standard2,000 Business credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Business 3000 creditsUSD 1,410/monthUSD 14,100Workspace · standard3,000 Business credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Business 4000 creditsUSD 1,840/monthUSD 18,400Workspace · standard4,000 Business credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Business 5000 creditsUSD 2,250/monthUSD 22,500Workspace · standard5,000 Business credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Business 7500 creditsUSD 3,300/monthUSD 33,000Workspace · standard7,500 Business credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Lovable / Business 10000 creditsUSD 4,300/monthUSD 43,000Workspace · standard10,000 Business credits/month + 5 build/day + Cloud/AI grants Pro starts at $25 for 100 credits; Business $50 for 100 differently valued credits, both per workspace with unlimited users. During the current rollout, general credits cover building, hosting and app AI after grants. All 11 published credit sizes per paid plan are captured. Equal credit counts across tiers are not equal dollar budgets. Unused monthly plan credits roll over while the subscription stays active, but they expire: two months from issue on monthly billing and one month after the annual billing period ends on annual billing, so an annual subscriber can bank a full year and a monthly one has a two-month window. Top-up credits last twelve months from purchase and referral credits twelve months from the grant; the daily 5 build credits never roll over (lovable.dev/pricing FAQ and docs.lovable.dev/introduction/plans-and-credits, read 2026-09-20). The earlier record that annual credits expire a year after issue and that unused months do not accumulate was wrong on both counts. Requests through connectors billed via Lovable (such as Perplexity, Firecrawl and Parallel) draw on the same credit balance (docs.lovable.dev/introduction/plans-and-credits, read 2026-09-27). If Cloud backend services pause after credits run out, replenishing credits may not restart them: the user may need More → Cloud → Overview → Wake up. Database/storage/authentication data remains retained but inaccessible while paused (plans-and-credits, read 2026-10-07). Paid workspaces can now set member, group, project and workspace credit limits and alerts; a hard limit blocks usage until an admin raises it or it resets. These optional controls do not increase the included credit pool (plans-and-credits, read 2026-10-08). The published temporary chat pricing and daily chat allowance ran through October 31, 2026. The credit guide still states that end date on November 4; no extension is established, so this promotion is not counted as current recurring capacity. Build and Plan mode, generated media, hosting and app AI retain standard pricing. The separately documented monthly Cloud/AI grants are 20/4 credits and remain temporary and subject to change. The current temporary grants also include 1 connector credit per month on Free, Pro and Business, consumed before general credits. This is not a permanent increase in the build allowance (official credits documentation, October 16). Publishing an app to invited people only allows up to ten people per project on Free and Pro. Business and Enterprise list this feature without that cap. Workspace membership and build credits are separate limits (official plan comparison, October 20). Credit top-ups now have a dedicated guide; Pro top-ups cost $15/50 credits and Business $30/50, valid for 12 months. Low-balance recovery does not add a published model-token conversion. All monthly tier selectors and the two-month annual discount were rechecked October 21. All 22 paid monthly and annual tiers were rechecked October 28 against rendered selectors and the exact yearly-charge documentation. Prices, credit amounts and rollover are unchanged. Newer models offered to deployed apps through the AI gateway are separate from the documented Opus 5 build agent; no model-token conversion is published for build credits.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3
Bolt / FreeFreeNot verifiedIndividual · standard300,000 tokens/day; 1 million/month Pro starts at $25/month or $216/year ($18/month equivalent); Teams starts at $30/member/month or $324/member/year ($27/month equivalent). Both start at 10 million vendor tokens/month with rollover; Free has 1 million/month and a 300,000/day limit. Token usage includes project context. The page does not identify the models and billing conversion needed to compare its internal token budget at API list prices. The expired Forge promotion is no longer an included-plan term (official rendered pricing page, October 20). Paid tokens expire two months after receipt on both monthly and annual billing while the paid subscription remains active. Teams tokens belong to each paid member and are not shared; free team members can view projects but cannot prompt (official pricing FAQ, November 4).Documented · checked 2026-11-03 Source 1
Bolt / Pro entry tierUSD 25/monthUSD 216Individual · startingStarts at 10 million tokens/month; rollover Pro starts at $25/month or $216/year ($18/month equivalent); Teams starts at $30/member/month or $324/member/year ($27/month equivalent). Both start at 10 million vendor tokens/month with rollover; Free has 1 million/month and a 300,000/day limit. Token usage includes project context. The page does not identify the models and billing conversion needed to compare its internal token budget at API list prices. The expired Forge promotion is no longer an included-plan term (official rendered pricing page, October 20). Paid tokens expire two months after receipt on both monthly and annual billing while the paid subscription remains active. Teams tokens belong to each paid member and are not shared; free team members can view projects but cannot prompt (official pricing FAQ, November 4).Documented · checked 2026-11-03 Source 1
Bolt / Teams entry tierUSD 30/monthUSD 324Team / seats · per memberTeam controls plus Pro features Pro starts at $25/month or $216/year ($18/month equivalent); Teams starts at $30/member/month or $324/member/year ($27/month equivalent). Both start at 10 million vendor tokens/month with rollover; Free has 1 million/month and a 300,000/day limit. Token usage includes project context. The page does not identify the models and billing conversion needed to compare its internal token budget at API list prices. The expired Forge promotion is no longer an included-plan term (official rendered pricing page, October 20). Paid tokens expire two months after receipt on both monthly and annual billing while the paid subscription remains active. Teams tokens belong to each paid member and are not shared; free team members can view projects but cannot prompt (official pricing FAQ, November 4).Documented · checked 2026-11-03 Source 1
v0 / FreeFreeNot verifiedIndividual · standard$5 one-time credits for new users; 7 messages/day v0.app/pricing changed the Free card from "$5 of included monthly credits" to "$5 in one-time credits for new users" (HTTP and browser read 2026-10-03), so Free no longer carries a recurring credit grant. The paid entry plan is now Plus. Plus and Business both include $30 monthly credit and $2/day on login: the higher subscription mainly buys controls, not more included inference. Names such as v0 Pro/Max are models, not subscription tiers. Hosting/runtime must be costed separately.Documented · checked 2026-11-02 Source 1
v0 / PlusUSD 30/monthNot verifiedIndividual · per user$30/month + $2 daily on login The paid entry plan is now Plus. Plus and Business both include $30 monthly credit and $2/day on login: the higher subscription mainly buys controls, not more included inference. Names such as v0 Pro/Max are models, not subscription tiers. Hosting/runtime must be costed separately.Documented · checked 2026-11-02 Source 1
v0 / BusinessUSD 100/monthNot verifiedIndividual · per user$30/month + $2 daily on login The paid entry plan is now Plus. Plus and Business both include $30 monthly credit and $2/day on login: the higher subscription mainly buys controls, not more included inference. Names such as v0 Pro/Max are models, not subscription tiers. Hosting/runtime must be costed separately.Documented · checked 2026-11-02 Source 1
OpenCode Go / GoUSD 10/monthNot verifiedIndividual · standardModel-specific usage-dollar limits; 5h/weekly/monthly caps One member per workspace; compatible coding clients required. The roster includes proprietary models (Grok 4.7, GPT 5.6 Luna) beside the open ones, each with its own monthly usage-dollar cap of $15-$100, of which the 5-hour window is 20% and the weekly window 50% (checked 2026-10-17). Kimi K2.7 Code joined the roster at a $60 monthly cap by 2026-09-21; it is also available as HighSpeed on qualifying native Kimi plans as of October 14. Muse Spark 1.3 and 1.2 Contributor, each with a $60 monthly cap, are discounted in exchange for permission to train on prompts and completions and are limited to regions permitted by Meta’s Geographic Use Policy. GLM-5.1, MiniMax M2.5, Qwen3.7 Max and Qwen3.6 Plus left the roster by 2026-09-29; The current DeepSeek V4.1 Flash monthly limit is $100 (raised from $80 by October 31); the five-hour and weekly fractions still apply. Model-specific monthly usage-dollar caps have five-hour 20% and weekly 50% caps. Request-count examples assume very high cache reuse; they are not guaranteed tasks. Supported coding-client traffic requirements apply. Do not sum model limits into one spendable cash balance or assume arbitrary API automation is included. OpenCode launched Go Plus at $40/month beside Go (opencode.ai/go and opencode.ai/docs/go, read 2026-09-29): same models and token prices, with per-model monthly caps two to eight times Go's (for example Kimi K3 $15 to $60, GLM-5.3 $15 to $120, DeepSeek V4.1 Flash $80 to $160 as verified October 17); it is a catalogue candidate rather than a graded offer. Step 5 Preview is now metered at $1.00/$2.70 per million input/output tokens and $0.10 cached input, with a base monthly cap of $15 on Go and $60 on Go Plus. The 4x Step 5 Preview promotion has ended; the recurring caps remain $15/$60. Five-hour and weekly controls still apply (official Go documentation, October 16). The October 17 model table removes Grok 4.6, Kimi K2.6 and MiniMax M2.7. Beam Free is temporary and region-limited; it is not recurring included access. The October 23 model table replaces MiniMax M3.1 Flash Free with metered MiniMax M3.1 Flash: $20 monthly usage on Go and $80 on Go Plus, at $0.15 input, $0.60 output and $0.015 cached input per million tokens. Hermes clients must use version 0.21.6 or later to send session headers for Go. When included use is exhausted, pay-as-you-go requires opting into Extra usage on the Overview credits card; it is not an automatic included-plan extension (official Go documentation, October 21). The October 22 roster and token-price table were reviewed in full. The temporary LongCat 2.5 Preview route has been removed; the remaining temporary Beam Free route does not contribute a recurring model option. Peak-hour DeepSeek and the stated short-context prices are used for the measured workload; larger contexts cost more. October24: Kimi K3.1 Pro joins Go with $15/month model usage ($60 on Go Plus), at $2 input/$10 output/$0.20 cache read/$2 cache write per million. All models consume shared five-hour/weekly/monthly allowances; these are alternative uses. Global routing is enabled by default only for requests from China; other workspaces must explicitly allow Global for DeepSeek. Contributor routes require the paid-training-endpoint setting. Temporary free offers remain excluded. October 25: Gemini 4 Argon now uses the standard gemini-4-argon route, with no training use and zero-day retention. Its recurring caps remain $30 on Go and $120 on Go Plus at standard $4 input/$20 output/$0.20 cache-read prices per million tokens. The 50% discount through November 30 is temporary and excluded from recurring capacity. Muse Contributor routes retain their separate training conditions. October 30: OpenCode adds Go Teams at $20 per member per month for workspaces with several members; each member gets their own limits, $5 a month above Go on every model, and Extra usage starts switched on. It is a team catalogue candidate, not a graded offer; Go and Go Plus remain single-member workspace plans. MiniMax M3 is now priced by context: $0.30 input, $1.20 output and $0.06 cached input per million up to 512K tokens and $0.60/$2.40/$0.12 above; the recorded 152K-token request stays in the lower band. October 31: DeepSeek V4.1 Flash monthly caps rise to $100 on Go, $200 on Go Plus and $105 on Go Teams (were $80/$160/$85) at unchanged peak and off-peak prices; the table's monthly request estimates rise accordingly. No promotion or end date is stated. November 5: Hy4 joins Go beside Hy4 preview at the same $0.834 input, $2.501 output and $0.042 cached-input prices per million tokens, with $60 monthly usage on Go, $240 on Go Plus and $65 on Go Teams; Hy4 preview keeps $30/$120/$35. Hy4 is not a ranked model here and is not substituted with Hy4 preview, so it adds no token option (official Go documentation, November 5).Documented · checked 2026-10-30 Source 1
Qoder / FreeFreeNot verifiedIndividual · standardLimited basic-model use; BYOK possible Monthly credits reset at term end. The current pricing page says users who continue on the Free plan receive basic models subject to daily and monthly limits; paid users can upgrade or buy credit packs. This does not establish an unlimited paid fallback (official pricing, October 22). Recurring base credits exclude temporary campaigns. Top-ups cost $20/1,500 credits, expire after one month, and are less favorable than base credit pricing. Fast replaces Performance across products from October 16, reducing its reference credit factor from 1.1x to 0.8x; Fast has context configuration but no thinking-effort setting. The IDE Quest window retired October 15 at 23:59 UTC+8. Editor Chat, Agent and Experts remain; task execution and mobile remote control migrate to the Qoder app using the same account, credits and benefits, with no extra purchase (official tier and retirement notices, read October 17). Lite is a new $10/month subscription with 1,000 recurring credits, full product features and a restricted model roster. Pro/Pro+/Ultra retain 2,000/6,000/20,000 credits at $20/$60/$200. October 21 first-purchase and renewal/upgrade bonuses are temporary, with no announced end date and 30-day expiry, and are excluded from recurring capacity. The event page gives a 1,000-credit Lite-to-Pro upgrade bonus while the launch article says 500; confirm the order-specific bonus at checkout. The general selector replaces MiniMax-M3 with MiniMax-M3.1-Flash, while the Lite announcement still names M3; Lite-specific MiniMax access is unresolved (official pricing, selector, Lite announcement and event terms, October 21). Sonus is available to individual Pro and higher subscribers, including after promotional credit packs expire by spending standard credits; Lite is excluded, and Free access requires an approved application. Community Boost runs October 22–30 (UTC+8); its welcome and round packs expire and are excluded from recurring capacity. Sonus has no verified matching ranked model or published credit-to-token conversion (official event terms, October 23). October 28: the general selector replaces Kimi K3 with Kimi K3.1 Pro at a 1.0× reference factor and removes Qwen3.7 Max and Plus. The Lite-specific list still names Kimi K3 and MiniMax M3, so the newer versions are not inferred for Lite. Context and reasoning controls depend on the selected model and client; no fixed effort menu is promised. Community Boost round-four packs expire on their respective October 28–30 release days and add no recurring capacity.Documented · checked 2026-10-30 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7 · Source 8
Qoder / LiteUSD 10/monthNot verifiedIndividual · standard1,000 credits/month Restricted model selection; full agent, Cloud Agents and Repo Wiki features. Temporary free-model and bonus-credit campaigns are not guaranteed recurring capacity. Monthly credits reset at term end. The current pricing page says users who continue on the Free plan receive basic models subject to daily and monthly limits; paid users can upgrade or buy credit packs. This does not establish an unlimited paid fallback (official pricing, October 22). Recurring base credits exclude temporary campaigns. Top-ups cost $20/1,500 credits, expire after one month, and are less favorable than base credit pricing. Fast replaces Performance across products from October 16, reducing its reference credit factor from 1.1x to 0.8x; Fast has context configuration but no thinking-effort setting. The IDE Quest window retired October 15 at 23:59 UTC+8. Editor Chat, Agent and Experts remain; task execution and mobile remote control migrate to the Qoder app using the same account, credits and benefits, with no extra purchase (official tier and retirement notices, read October 17). Lite is a new $10/month subscription with 1,000 recurring credits, full product features and a restricted model roster. Pro/Pro+/Ultra retain 2,000/6,000/20,000 credits at $20/$60/$200. October 21 first-purchase and renewal/upgrade bonuses are temporary, with no announced end date and 30-day expiry, and are excluded from recurring capacity. The event page gives a 1,000-credit Lite-to-Pro upgrade bonus while the launch article says 500; confirm the order-specific bonus at checkout. The general selector replaces MiniMax-M3 with MiniMax-M3.1-Flash, while the Lite announcement still names M3; Lite-specific MiniMax access is unresolved (official pricing, selector, Lite announcement and event terms, October 21). Sonus is available to individual Pro and higher subscribers, including after promotional credit packs expire by spending standard credits; Lite is excluded, and Free access requires an approved application. Community Boost runs October 22–30 (UTC+8); its welcome and round packs expire and are excluded from recurring capacity. Sonus has no verified matching ranked model or published credit-to-token conversion (official event terms, October 23). October 28: the general selector replaces Kimi K3 with Kimi K3.1 Pro at a 1.0× reference factor and removes Qwen3.7 Max and Plus. The Lite-specific list still names Kimi K3 and MiniMax M3, so the newer versions are not inferred for Lite. Context and reasoning controls depend on the selected model and client; no fixed effort menu is promised. Community Boost round-four packs expire on their respective October 28–30 release days and add no recurring capacity.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7 · Source 8
Qoder / ProUSD 20/monthNot verifiedIndividual · standard2,000 credits/month Monthly credits reset at term end. The current pricing page says users who continue on the Free plan receive basic models subject to daily and monthly limits; paid users can upgrade or buy credit packs. This does not establish an unlimited paid fallback (official pricing, October 22). Recurring base credits exclude temporary campaigns. Top-ups cost $20/1,500 credits, expire after one month, and are less favorable than base credit pricing. Fast replaces Performance across products from October 16, reducing its reference credit factor from 1.1x to 0.8x; Fast has context configuration but no thinking-effort setting. The IDE Quest window retired October 15 at 23:59 UTC+8. Editor Chat, Agent and Experts remain; task execution and mobile remote control migrate to the Qoder app using the same account, credits and benefits, with no extra purchase (official tier and retirement notices, read October 17). Lite is a new $10/month subscription with 1,000 recurring credits, full product features and a restricted model roster. Pro/Pro+/Ultra retain 2,000/6,000/20,000 credits at $20/$60/$200. October 21 first-purchase and renewal/upgrade bonuses are temporary, with no announced end date and 30-day expiry, and are excluded from recurring capacity. The event page gives a 1,000-credit Lite-to-Pro upgrade bonus while the launch article says 500; confirm the order-specific bonus at checkout. The general selector replaces MiniMax-M3 with MiniMax-M3.1-Flash, while the Lite announcement still names M3; Lite-specific MiniMax access is unresolved (official pricing, selector, Lite announcement and event terms, October 21). Sonus is available to individual Pro and higher subscribers, including after promotional credit packs expire by spending standard credits; Lite is excluded, and Free access requires an approved application. Community Boost runs October 22–30 (UTC+8); its welcome and round packs expire and are excluded from recurring capacity. Sonus has no verified matching ranked model or published credit-to-token conversion (official event terms, October 23). October 28: the general selector replaces Kimi K3 with Kimi K3.1 Pro at a 1.0× reference factor and removes Qwen3.7 Max and Plus. The Lite-specific list still names Kimi K3 and MiniMax M3, so the newer versions are not inferred for Lite. Context and reasoning controls depend on the selected model and client; no fixed effort menu is promised. Community Boost round-four packs expire on their respective October 28–30 release days and add no recurring capacity.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7 · Source 8
Qoder / Pro+USD 60/monthNot verifiedIndividual · standard6,000 credits/month Monthly credits reset at term end. The current pricing page says users who continue on the Free plan receive basic models subject to daily and monthly limits; paid users can upgrade or buy credit packs. This does not establish an unlimited paid fallback (official pricing, October 22). Recurring base credits exclude temporary campaigns. Top-ups cost $20/1,500 credits, expire after one month, and are less favorable than base credit pricing. Fast replaces Performance across products from October 16, reducing its reference credit factor from 1.1x to 0.8x; Fast has context configuration but no thinking-effort setting. The IDE Quest window retired October 15 at 23:59 UTC+8. Editor Chat, Agent and Experts remain; task execution and mobile remote control migrate to the Qoder app using the same account, credits and benefits, with no extra purchase (official tier and retirement notices, read October 17). Lite is a new $10/month subscription with 1,000 recurring credits, full product features and a restricted model roster. Pro/Pro+/Ultra retain 2,000/6,000/20,000 credits at $20/$60/$200. October 21 first-purchase and renewal/upgrade bonuses are temporary, with no announced end date and 30-day expiry, and are excluded from recurring capacity. The event page gives a 1,000-credit Lite-to-Pro upgrade bonus while the launch article says 500; confirm the order-specific bonus at checkout. The general selector replaces MiniMax-M3 with MiniMax-M3.1-Flash, while the Lite announcement still names M3; Lite-specific MiniMax access is unresolved (official pricing, selector, Lite announcement and event terms, October 21). Sonus is available to individual Pro and higher subscribers, including after promotional credit packs expire by spending standard credits; Lite is excluded, and Free access requires an approved application. Community Boost runs October 22–30 (UTC+8); its welcome and round packs expire and are excluded from recurring capacity. Sonus has no verified matching ranked model or published credit-to-token conversion (official event terms, October 23). October 28: the general selector replaces Kimi K3 with Kimi K3.1 Pro at a 1.0× reference factor and removes Qwen3.7 Max and Plus. The Lite-specific list still names Kimi K3 and MiniMax M3, so the newer versions are not inferred for Lite. Context and reasoning controls depend on the selected model and client; no fixed effort menu is promised. Community Boost round-four packs expire on their respective October 28–30 release days and add no recurring capacity.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7 · Source 8
Qoder / UltraUSD 200/monthNot verifiedIndividual · standard20,000 credits/month Monthly credits reset at term end. The current pricing page says users who continue on the Free plan receive basic models subject to daily and monthly limits; paid users can upgrade or buy credit packs. This does not establish an unlimited paid fallback (official pricing, October 22). Recurring base credits exclude temporary campaigns. Top-ups cost $20/1,500 credits, expire after one month, and are less favorable than base credit pricing. Fast replaces Performance across products from October 16, reducing its reference credit factor from 1.1x to 0.8x; Fast has context configuration but no thinking-effort setting. The IDE Quest window retired October 15 at 23:59 UTC+8. Editor Chat, Agent and Experts remain; task execution and mobile remote control migrate to the Qoder app using the same account, credits and benefits, with no extra purchase (official tier and retirement notices, read October 17). Lite is a new $10/month subscription with 1,000 recurring credits, full product features and a restricted model roster. Pro/Pro+/Ultra retain 2,000/6,000/20,000 credits at $20/$60/$200. October 21 first-purchase and renewal/upgrade bonuses are temporary, with no announced end date and 30-day expiry, and are excluded from recurring capacity. The event page gives a 1,000-credit Lite-to-Pro upgrade bonus while the launch article says 500; confirm the order-specific bonus at checkout. The general selector replaces MiniMax-M3 with MiniMax-M3.1-Flash, while the Lite announcement still names M3; Lite-specific MiniMax access is unresolved (official pricing, selector, Lite announcement and event terms, October 21). Sonus is available to individual Pro and higher subscribers, including after promotional credit packs expire by spending standard credits; Lite is excluded, and Free access requires an approved application. Community Boost runs October 22–30 (UTC+8); its welcome and round packs expire and are excluded from recurring capacity. Sonus has no verified matching ranked model or published credit-to-token conversion (official event terms, October 23). October 28: the general selector replaces Kimi K3 with Kimi K3.1 Pro at a 1.0× reference factor and removes Qwen3.7 Max and Plus. The Lite-specific list still names Kimi K3 and MiniMax M3, so the newer versions are not inferred for Lite. Context and reasoning controls depend on the selected model and client; no fixed effort menu is promised. Community Boost round-four packs expire on their respective October 28–30 release days and add no recurring capacity.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7 · Source 8
Tencent CodeBuddy / FreeFreeNot verifiedIndividual · standard100 credits/month; 5,000 completions/month Pro includes 1,000 base plus 1,000 monthly bonus credits under the current terms. Daily activity bonuses are separate temporary promotions. Completions and automated tasks have their own limits; team credits are shared. A limited-time daily activity reward adds 30 credits a day on Free and 50 on Pro for any day the user sends a chat request; it is promotional and is not scored as recurring capacity (codebuddy.ai docs pricing, read 2026-09-20). Base and bonus credits are valid for the month they are issued and do not roll over.Documented · checked 2026-11-02 Source 1
Tencent CodeBuddy / ProUSD 10/monthUSD 96Individual · standard1,000 base + 1,000 bonus credits/month Pro includes 1,000 base plus 1,000 monthly bonus credits under the current terms. Daily activity bonuses are separate temporary promotions. Completions and automated tasks have their own limits; team credits are shared. A limited-time daily activity reward adds 30 credits a day on Free and 50 on Pro for any day the user sends a chat request; it is promotional and is not scored as recurring capacity (codebuddy.ai docs pricing, read 2026-09-20). Base and bonus credits are valid for the month they are issued and do not roll over.Documented · checked 2026-11-02 Source 1
Tencent CodeBuddy / TeamUSD 40/monthUSD 480Team / seats · per seat1,000 credits/seat/month pooled Pro includes 1,000 base plus 1,000 monthly bonus credits under the current terms. Daily activity bonuses are separate temporary promotions. Completions and automated tasks have their own limits; team credits are shared. A limited-time daily activity reward adds 30 credits a day on Free and 50 on Pro for any day the user sends a chat request; it is promotional and is not scored as recurring capacity (codebuddy.ai docs pricing, read 2026-09-20). Base and bonus credits are valid for the month they are issued and do not roll over.Documented · checked 2026-11-02 Source 1
Blackbox / Enterprise inferencePrice not verifiedNot verifiedEnterprise · custom annual commitmentCustom committed spend and metered tokens; no public price. Current pricing page describes custom annual enterprise inference commitments, metered by token, with no public consumer flat-rate plan. Historical consumer membership prices were not imported. Treat as an enterprise/API alternative, not a $/month individual contender. November 2 correction: the enterprise card shows "Closed −5%+ Open −10%+" beside committed spend (in every capture since October 27), and the page states the same rack rates with no markups, metering commits at the published per-model rates; the October 28 note that it no longer promised discounts was wrong. The commitment is contract-specific; no public recurring consumer charge is established.Usage-based / enterprise · checked 2026-11-02 Source 1
TRAE / FreeFreeNot verifiedIndividual · standardAuto mode only; limited usage; 5,000 completions/month; 2 concurrent cloud tasks Monthly prices and usage values are public. Pro trial is excluded. The pricing card and the comparison table now state the same dollar figure for every paid tier, and the card's multipliers (3× Pro, 10× Pro) agree with those figures. Dollar-labelled usage follows TRAE billing rules, not cash reimbursement. October 31: TRAE's rendered model page adds GPT-6.1 Sol to the models of the current plans, at $2 input, $0.10 cache read, $2.50 cache write and $10 output per million tokens up to 272K ($4/$0.20/$5/$15 above) in both regional tables. GPT-6.1 Terra, DeepSeek V4.1 Flash and DeepSeek V4 Pro appear only in the non-U.S. table and are not added to the U.S.-dollar plan rosters. November 5: after TraeCode and TraeWork merged into TRAE, the rendered model page lists no Claude model in either regional table. The U.S.-dollar plan rosters never included Claude, and the GPT-6 Astra, GPT-6.1 Sol, GPT-6 Sol and GPT-5.6 Sol rates are unchanged (official model page, browser-read November 5).Documented · checked 2026-10-30 Source 1
TRAE / LitePrice not verifiedNot verifiedIndividual · standardRemoved from the pricing page on or before 2026-09-17 and still absent on 2026-09-18; the paid tiers are now Pro, Pro+ and Ultra. Monthly prices and usage values are public. Pro trial is excluded. The pricing card and the comparison table now state the same dollar figure for every paid tier, and the card's multipliers (3× Pro, 10× Pro) agree with those figures. Dollar-labelled usage follows TRAE billing rules, not cash reimbursement. October 31: TRAE's rendered model page adds GPT-6.1 Sol to the models of the current plans, at $2 input, $0.10 cache read, $2.50 cache write and $10 output per million tokens up to 272K ($4/$0.20/$5/$15 above) in both regional tables. GPT-6.1 Terra, DeepSeek V4.1 Flash and DeepSeek V4 Pro appear only in the non-U.S. table and are not added to the U.S.-dollar plan rosters. November 5: after TraeCode and TraeWork merged into TRAE, the rendered model page lists no Claude model in either regional table. The U.S.-dollar plan rosters never included Claude, and the GPT-6 Astra, GPT-6.1 Sol, GPT-6 Sol and GPT-5.6 Sol rates are unchanged (official model page, browser-read November 5).Retired · checked 2026-10-30 Source 1
TRAE / ProUSD 20/monthNot verifiedIndividual · standard$20 basic usage; unlimited autocomplete; up to 10 concurrent cloud tasks Monthly prices and usage values are public. Pro trial is excluded. The pricing card and the comparison table now state the same dollar figure for every paid tier, and the card's multipliers (3× Pro, 10× Pro) agree with those figures. Dollar-labelled usage follows TRAE billing rules, not cash reimbursement. October 31: TRAE's rendered model page adds GPT-6.1 Sol to the models of the current plans, at $2 input, $0.10 cache read, $2.50 cache write and $10 output per million tokens up to 272K ($4/$0.20/$5/$15 above) in both regional tables. GPT-6.1 Terra, DeepSeek V4.1 Flash and DeepSeek V4 Pro appear only in the non-U.S. table and are not added to the U.S.-dollar plan rosters. November 5: after TraeCode and TraeWork merged into TRAE, the rendered model page lists no Claude model in either regional table. The U.S.-dollar plan rosters never included Claude, and the GPT-6 Astra, GPT-6.1 Sol, GPT-6 Sol and GPT-5.6 Sol rates are unchanged (official model page, browser-read November 5).Documented · checked 2026-10-30 Source 1
TRAE / Pro+USD 60/monthNot verifiedIndividual · standard$60 basic usage (3× Pro); up to 15 concurrent cloud tasks Monthly prices and usage values are public. Pro trial is excluded. The pricing card and the comparison table now state the same dollar figure for every paid tier, and the card's multipliers (3× Pro, 10× Pro) agree with those figures. Dollar-labelled usage follows TRAE billing rules, not cash reimbursement. October 31: TRAE's rendered model page adds GPT-6.1 Sol to the models of the current plans, at $2 input, $0.10 cache read, $2.50 cache write and $10 output per million tokens up to 272K ($4/$0.20/$5/$15 above) in both regional tables. GPT-6.1 Terra, DeepSeek V4.1 Flash and DeepSeek V4 Pro appear only in the non-U.S. table and are not added to the U.S.-dollar plan rosters. November 5: after TraeCode and TraeWork merged into TRAE, the rendered model page lists no Claude model in either regional table. The U.S.-dollar plan rosters never included Claude, and the GPT-6 Astra, GPT-6.1 Sol, GPT-6 Sol and GPT-5.6 Sol rates are unchanged (official model page, browser-read November 5).Documented · checked 2026-10-30 Source 1
TRAE / UltraUSD 200/monthNot verifiedIndividual · standard$200 basic usage (10× Pro); model early access; up to 20 concurrent cloud tasks Monthly prices and usage values are public. Pro trial is excluded. The pricing card and the comparison table now state the same dollar figure for every paid tier, and the card's multipliers (3× Pro, 10× Pro) agree with those figures. Dollar-labelled usage follows TRAE billing rules, not cash reimbursement. October 31: TRAE's rendered model page adds GPT-6.1 Sol to the models of the current plans, at $2 input, $0.10 cache read, $2.50 cache write and $10 output per million tokens up to 272K ($4/$0.20/$5/$15 above) in both regional tables. GPT-6.1 Terra, DeepSeek V4.1 Flash and DeepSeek V4 Pro appear only in the non-U.S. table and are not added to the U.S.-dollar plan rosters. November 5: after TraeCode and TraeWork merged into TRAE, the rendered model page lists no Claude model in either regional table. The U.S.-dollar plan rosters never included Claude, and the GPT-6 Astra, GPT-6.1 Sol, GPT-6 Sol and GPT-5.6 Sol rates are unchanged (official model page, browser-read November 5).Documented · checked 2026-10-30 Source 1
Cline / Open source + inferenceFree software · inference extraNot verifiedUsage-based · software fee onlyInference paid separately or supplied by user Current individual offering is free software plus separately paid inference/BYOK, explicitly without a subscription. A zero software price is not zero total AI spend. Enterprise has custom governance pricing. ClinePass (cline.bot/cline-pass, read 2026-09-25) is a separate $9.99/month subscription for a curated list of open-weight models from six Chinese labs with unquantified quotas described as 2-5x standard API rate limits; the page calls $9.99 the standard rate after a promotion period and says an additional processing fee may apply, so it is a catalogue candidate rather than a graded offer.Usage-based / enterprise · checked 2026-11-02 Source 1
Kilo / Kilo Pass / Pass StarterUSD 19.95/monthNot verifiedIndividual · standard$19 base paid credits + conditional bonus $19 Pass plus the 5% processing fee. Pass converts subscription dollars into paid credits. Monthly bonus depends on tenure (10% in month 1, 50% welcome bonus in month 2, then a growing streak bonus reaching 40% in month 8); annual offers 50% monthly. Bonus credits expire monthly. A 5% processing fee applies to monthly and annual Pass purchases and adds no credits (pricing FAQ: "a $20 credit purchase costs $21", read 2026-10-19). Team software fees and cloud compute are separate from inference. Partner coding plans must not be double-counted. Kilo announces on its pricing page that it has been acquired by Anaconda.Documented · checked 2026-11-02 Source 1 · Source 2
Kilo / Kilo Pass / Pass ProUSD 51.45/monthNot verifiedIndividual · standard$49 base paid credits + conditional bonus $49 Pass plus the 5% processing fee. Pass converts subscription dollars into paid credits. Monthly bonus depends on tenure (10% in month 1, 50% welcome bonus in month 2, then a growing streak bonus reaching 40% in month 8); annual offers 50% monthly. Bonus credits expire monthly. A 5% processing fee applies to monthly and annual Pass purchases and adds no credits (pricing FAQ: "a $20 credit purchase costs $21", read 2026-10-19). Team software fees and cloud compute are separate from inference. Partner coding plans must not be double-counted. Kilo announces on its pricing page that it has been acquired by Anaconda.Documented · checked 2026-11-02 Source 1 · Source 2
Kilo / Kilo Pass / Pass ExpertUSD 208.95/monthNot verifiedIndividual · standard$199 base paid credits + conditional bonus $199 Pass plus the 5% processing fee. Pass converts subscription dollars into paid credits. Monthly bonus depends on tenure (10% in month 1, 50% welcome bonus in month 2, then a growing streak bonus reaching 40% in month 8); annual offers 50% monthly. Bonus credits expire monthly. A 5% processing fee applies to monthly and annual Pass purchases and adds no credits (pricing FAQ: "a $20 credit purchase costs $21", read 2026-10-19). Team software fees and cloud compute are separate from inference. Partner coding plans must not be double-counted. Kilo announces on its pricing page that it has been acquired by Anaconda.Documented · checked 2026-11-02 Source 1 · Source 2
Kilo / Kilo Pass / Teams platformUSD 15/monthNot verifiedTeam / seats · per user + usageNo included inference established Separate from Pass/inference and cloud compute. Pass converts subscription dollars into paid credits. Monthly bonus depends on tenure (10% in month 1, 50% welcome bonus in month 2, then a growing streak bonus reaching 40% in month 8); annual offers 50% monthly. Bonus credits expire monthly. A 5% processing fee applies to monthly and annual Pass purchases and adds no credits (pricing FAQ: "a $20 credit purchase costs $21", read 2026-10-19). Team software fees and cloud compute are separate from inference. Partner coding plans must not be double-counted. Kilo announces on its pricing page that it has been acquired by Anaconda.Documented · checked 2026-10-31 Source 1 · Source 2
CommandCode / GoUSD 1/monthNot verifiedIndividual · base + unspecified mandatory fee$10 monthly pool, moving from a flat wallet to per-model monthly credits at the next billing cycle; windows given as $2 per 5 hours and $5 per week in the limits tables but $3 and $6 in the migration FAQ; no API access on this tier. Reaching a shorter window can pause included usage before monthly credits run out. Optional paid credits can continue work. Mandatory processing fee is not included in the listed price. Advertised credit multiples are not guaranteed by the terms; additional credits depend on availability. Advertised monthly credits have five-hour, weekly and per-model limits. On the Max plans premium and standard models each carry their own limit ($100 premium / $150 standard on Max 10x, $200 / $300 on Max 20x), so the headline credit figure is not one spendable pool. The Go plan is the only plan without API access, and some models are geo-restricted by their providers (GPT-5.6 Luna and Gemini 3.7 Flash are unavailable in China). Per-model tier gating applies: several models are documented as available on GOAT and above only. The terms, reissued October 20, 2026 (read October 23), now state the ceiling outright: subscriptions are sold on the basis of underlying API cost and a buyer "may expect, at maximum, a one-to-one (1:1) ratio between the subscription amount paid and the API credits made available", with no guarantee of any greater ratio and actual usage possibly lower after model pricing, provider rate changes, taxes and overhead. Every advertised allowance above 1x the price - Go at $1 against a $10 usage reference, GOAT 70 credits for $10, Pro 80 for $20, both Max tiers at 1.5x - is therefore a marketing figure the terms disclaim. Additional, bonus, promotional and top-up credits are "while supplies last" and revocable at any time, and unused credits and capacity do not roll over to a later period. Coding-plan processing fees remain unspecified. Provider documentation gives a $1.01 processing fee (called a card fee until 2026-10-01) but conflicting recurring versus pay-as-you-go wording; plans and extra credits can be paid with a card, Alipay, WeChat Pay or crypto in USD, or UPI in INR (official payment methods read 2026-11-03). The terms (last updated October 20, 2026; read 2026-10-23) also allow one account per individual and reserve a permanent ban for multi-account use. The Go plan moves from a flat $10 wallet to per-model monthly credits (one shared $10 pool with larger allowances on selected models) at each subscriber's next billing cycle; the pricing-limits and usage-limits tables now give Go windows of $2 per 5 hours and $5 per week while the same page's migration FAQ says they stay at $3 and $6, its request estimate fell from ~15K to ~9K, and DeepSeek V4.1 Flash is boosted to the full $10 on Go (commandcode.ai/docs/resources/pricing-limits, read 2026-09-29). Mistral Large 4 is now listed across plans, including Go, with temporary boosted per-model monthly limits of $10/$40/$50 on Go/GOAT/Pro. These promotional boosts do not resolve the existing terms-versus-allowance conflict. Cloud agent access is available on every plan except Go, billed per second at $0.10/hour; it is not an included runtime allowance (pricing, read 2026-10-08). Claude Haiku 5.5 is now also available on GOAT at the $20 default per-model allowance (pricing-page release notes, read 2026-10-11); Pro and above retain access. The per-model allowance does not override the conflicting plan-wide terms. Glyph Cluster is a temporary free stealth preview limited to 1,000 requests/day (100 before October 31) with at least $1 credit balance; it is not a recurring allowance. The GLM-5.3 Flash 100-credit promotion ended October 17; GLM-5.3 Flash now has boosted per-model allowances of $10/$80/$90 on Go/GOAT/Pro "while capacity lasts" at unchanged token rates (official pricing-limits and provider docs, October 18). These model-specific figures do not resolve the separate contractual 1:1 ceiling, and do not increase our capacity assessment (October 15). Ling 3.1 Flash now has boosted per-model credits of $100 on GOAT and $110 on Pro at unchanged token rates. The temporary preview does not establish recurring included capacity. The DeepSeek V4.1 Flash temporary 100-credit promotion ended October 16; the October 17 per-model table showed $10/$80/$90 on Go/GOAT/Pro; the October 20 limits page no longer establishes these as a permanent boost. Beam 1.0 is a temporary free route while capacity lasts, with a $1 balance requirement and possible training on traffic; it is not a recurring subscription entitlement (official pricing and provider docs, October 17). The October 20 pricing page advertises 80 GOAT credits while its comparison table still says 70, and the terms retain the 1:1 API-cost ceiling. That unresolved conflict prevents a capacity figure. MiniMax M3.1 Flash Free is a temporary ZDR route with no daily cap and a $1 balance requirement, not a recurring token grant. The current provider table no longer publishes per-model ZDR check marks; historical symbols do not establish current retention terms. Ornith 1.5 122B Free is temporary through October 21 at 11 PM Pacific and requires a $1 balance. The DeepSeek October17 launch boost expired October21 at 01:00 UTC. Temporary rates do not establish a recurring allowance or resolve the GOAT 80-versus-70-credit and contractual 1:1-ceiling conflict (official provider, limits and terms, October 21). The October 22 usage page removes the expired Mistral promotion and shows a new temporary DeepSeek boost; neither adds recurring capacity. The GOAT price/credit and contractual ceiling discrepancies remain unresolved. The October 23 pages remove the MiMo V2.5/V2.5 Pro offers and expired MiniMax M3.1 Flash/Ornith free promotions. Paid MiniMax and Ornith routes remain, but no recurring capacity figure is established while the contractual and advertised credit totals conflict. October24: expired free TypeSafe/Ornith promotions are absent; GOAT100 listsLing3.1Flash, while the Pro300Qwen3.7Max offer is removed. The advertised balances still conflict with the contractual1:1 ceiling; no graded recurring allowance is assigned. October 27: the limits tables now advertise up to $100 GOAT and $110 Pro, with up to $20/$50 and $22/$55 five-hour/weekly limits respectively. These maxima correspond to Ling 3.1 Flash; GLM-5.2 remains the $70/$80 example. The October 28 pricing card now says up to $100 GOAT, and its comparison table agrees on up to $100 GOAT/$110 Pro. Mandatory processing charges and the contractual 1:1 API-cost ceiling remain unresolved, so no recurring token value is assigned. GPT-6.1 Terra is now listed by the Provider rate card; per-plan availability is recorded only where explicitly verified. October 28: current terms are dated October 21 and remove the assurance that CommandCode does not train its own models on submitted material. Pricing and provider pages remove the matching no-training statement; zero-retention enforcement remains conditional on provider support. These are revised disclosures, not evidence that training occurred. Mandatory fees and the contractual 1:1 API-cost ceiling still prevent a verified recurring token figure. October 31: DeepSeek V4.1 Flash and GLM-5.3 Flash boosted per-model allowances are $10/$100/$110 on Go/GOAT/Pro (were $10/$80/$90); the DeepSeek deal no longer says it has no end date. The Ling 3.1 Flash 1019 free promotion ended November 3. Crouton Alpha is now a temporary stealth preview, capped at 1,000 requests a day while it lasts, with a $1 balance requirement and possible provider retention of prompts and outputs. It is not a recurring allowance (official pricing and limits, November 4). Fees and the contractual 1:1 ceiling still prevent a figure.Conditions apply · checked 2026-10-31 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
CommandCode / GOATUSD 10/monthNot verifiedIndividual · base + unspecified mandatory feeUp to $100 monthly model usage; up to $20 per 5-hour window and $50 per weekly window. Model-specific allowances vary; these maxima use Ling 3.1 Flash. Reaching a shorter window can pause included usage before monthly credits run out. Optional paid credits can continue work. Mandatory processing fee is not included in the listed price. Advertised credit multiples are not guaranteed by the terms; additional credits depend on availability. Advertised monthly credits have five-hour, weekly and per-model limits. On the Max plans premium and standard models each carry their own limit ($100 premium / $150 standard on Max 10x, $200 / $300 on Max 20x), so the headline credit figure is not one spendable pool. The Go plan is the only plan without API access, and some models are geo-restricted by their providers (GPT-5.6 Luna and Gemini 3.7 Flash are unavailable in China). Per-model tier gating applies: several models are documented as available on GOAT and above only. The terms, reissued October 20, 2026 (read October 23), now state the ceiling outright: subscriptions are sold on the basis of underlying API cost and a buyer "may expect, at maximum, a one-to-one (1:1) ratio between the subscription amount paid and the API credits made available", with no guarantee of any greater ratio and actual usage possibly lower after model pricing, provider rate changes, taxes and overhead. Every advertised allowance above 1x the price - Go at $1 against a $10 usage reference, GOAT 70 credits for $10, Pro 80 for $20, both Max tiers at 1.5x - is therefore a marketing figure the terms disclaim. Additional, bonus, promotional and top-up credits are "while supplies last" and revocable at any time, and unused credits and capacity do not roll over to a later period. Coding-plan processing fees remain unspecified. Provider documentation gives a $1.01 processing fee (called a card fee until 2026-10-01) but conflicting recurring versus pay-as-you-go wording; plans and extra credits can be paid with a card, Alipay, WeChat Pay or crypto in USD, or UPI in INR (official payment methods read 2026-11-03). The terms (last updated October 20, 2026; read 2026-10-23) also allow one account per individual and reserve a permanent ban for multi-account use. The Go plan moves from a flat $10 wallet to per-model monthly credits (one shared $10 pool with larger allowances on selected models) at each subscriber's next billing cycle; the pricing-limits and usage-limits tables now give Go windows of $2 per 5 hours and $5 per week while the same page's migration FAQ says they stay at $3 and $6, its request estimate fell from ~15K to ~9K, and DeepSeek V4.1 Flash is boosted to the full $10 on Go (commandcode.ai/docs/resources/pricing-limits, read 2026-09-29). Mistral Large 4 is now listed across plans, including Go, with temporary boosted per-model monthly limits of $10/$40/$50 on Go/GOAT/Pro. These promotional boosts do not resolve the existing terms-versus-allowance conflict. Cloud agent access is available on every plan except Go, billed per second at $0.10/hour; it is not an included runtime allowance (pricing, read 2026-10-08). Claude Haiku 5.5 is now also available on GOAT at the $20 default per-model allowance (pricing-page release notes, read 2026-10-11); Pro and above retain access. The per-model allowance does not override the conflicting plan-wide terms. Glyph Cluster is a temporary free stealth preview limited to 1,000 requests/day (100 before October 31) with at least $1 credit balance; it is not a recurring allowance. The GLM-5.3 Flash 100-credit promotion ended October 17; GLM-5.3 Flash now has boosted per-model allowances of $10/$80/$90 on Go/GOAT/Pro "while capacity lasts" at unchanged token rates (official pricing-limits and provider docs, October 18). These model-specific figures do not resolve the separate contractual 1:1 ceiling, and do not increase our capacity assessment (October 15). Ling 3.1 Flash now has boosted per-model credits of $100 on GOAT and $110 on Pro at unchanged token rates. The temporary preview does not establish recurring included capacity. The DeepSeek V4.1 Flash temporary 100-credit promotion ended October 16; the October 17 per-model table showed $10/$80/$90 on Go/GOAT/Pro; the October 20 limits page no longer establishes these as a permanent boost. Beam 1.0 is a temporary free route while capacity lasts, with a $1 balance requirement and possible training on traffic; it is not a recurring subscription entitlement (official pricing and provider docs, October 17). The October 20 pricing page advertises 80 GOAT credits while its comparison table still says 70, and the terms retain the 1:1 API-cost ceiling. That unresolved conflict prevents a capacity figure. MiniMax M3.1 Flash Free is a temporary ZDR route with no daily cap and a $1 balance requirement, not a recurring token grant. The current provider table no longer publishes per-model ZDR check marks; historical symbols do not establish current retention terms. Ornith 1.5 122B Free is temporary through October 21 at 11 PM Pacific and requires a $1 balance. The DeepSeek October17 launch boost expired October21 at 01:00 UTC. Temporary rates do not establish a recurring allowance or resolve the GOAT 80-versus-70-credit and contractual 1:1-ceiling conflict (official provider, limits and terms, October 21). The October 22 usage page removes the expired Mistral promotion and shows a new temporary DeepSeek boost; neither adds recurring capacity. The GOAT price/credit and contractual ceiling discrepancies remain unresolved. The October 23 pages remove the MiMo V2.5/V2.5 Pro offers and expired MiniMax M3.1 Flash/Ornith free promotions. Paid MiniMax and Ornith routes remain, but no recurring capacity figure is established while the contractual and advertised credit totals conflict. October24: expired free TypeSafe/Ornith promotions are absent; GOAT100 listsLing3.1Flash, while the Pro300Qwen3.7Max offer is removed. The advertised balances still conflict with the contractual1:1 ceiling; no graded recurring allowance is assigned. October 27: the limits tables now advertise up to $100 GOAT and $110 Pro, with up to $20/$50 and $22/$55 five-hour/weekly limits respectively. These maxima correspond to Ling 3.1 Flash; GLM-5.2 remains the $70/$80 example. The October 28 pricing card now says up to $100 GOAT, and its comparison table agrees on up to $100 GOAT/$110 Pro. Mandatory processing charges and the contractual 1:1 API-cost ceiling remain unresolved, so no recurring token value is assigned. GPT-6.1 Terra is now listed by the Provider rate card; per-plan availability is recorded only where explicitly verified. October 28: current terms are dated October 21 and remove the assurance that CommandCode does not train its own models on submitted material. Pricing and provider pages remove the matching no-training statement; zero-retention enforcement remains conditional on provider support. These are revised disclosures, not evidence that training occurred. Mandatory fees and the contractual 1:1 API-cost ceiling still prevent a verified recurring token figure. October 31: DeepSeek V4.1 Flash and GLM-5.3 Flash boosted per-model allowances are $10/$100/$110 on Go/GOAT/Pro (were $10/$80/$90); the DeepSeek deal no longer says it has no end date. The Ling 3.1 Flash 1019 free promotion ended November 3. Crouton Alpha is now a temporary stealth preview, capped at 1,000 requests a day while it lasts, with a $1 balance requirement and possible provider retention of prompts and outputs. It is not a recurring allowance (official pricing and limits, November 4). Fees and the contractual 1:1 ceiling still prevent a figure.Conditions apply · checked 2026-10-31 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
CommandCode / ProUSD 20/monthNot verifiedIndividual · base + unspecified mandatory feeUp to $110 monthly model usage; up to $22 per 5-hour window and $55 per weekly window. Model-specific allowances vary; these maxima use Ling 3.1 Flash. Reaching a shorter window can pause included usage before monthly credits run out. Optional paid credits can continue work. Mandatory processing fee is not included in the listed price. Advertised credit multiples are not guaranteed by the terms; additional credits depend on availability. Advertised monthly credits have five-hour, weekly and per-model limits. On the Max plans premium and standard models each carry their own limit ($100 premium / $150 standard on Max 10x, $200 / $300 on Max 20x), so the headline credit figure is not one spendable pool. The Go plan is the only plan without API access, and some models are geo-restricted by their providers (GPT-5.6 Luna and Gemini 3.7 Flash are unavailable in China). Per-model tier gating applies: several models are documented as available on GOAT and above only. The terms, reissued October 20, 2026 (read October 23), now state the ceiling outright: subscriptions are sold on the basis of underlying API cost and a buyer "may expect, at maximum, a one-to-one (1:1) ratio between the subscription amount paid and the API credits made available", with no guarantee of any greater ratio and actual usage possibly lower after model pricing, provider rate changes, taxes and overhead. Every advertised allowance above 1x the price - Go at $1 against a $10 usage reference, GOAT 70 credits for $10, Pro 80 for $20, both Max tiers at 1.5x - is therefore a marketing figure the terms disclaim. Additional, bonus, promotional and top-up credits are "while supplies last" and revocable at any time, and unused credits and capacity do not roll over to a later period. Coding-plan processing fees remain unspecified. Provider documentation gives a $1.01 processing fee (called a card fee until 2026-10-01) but conflicting recurring versus pay-as-you-go wording; plans and extra credits can be paid with a card, Alipay, WeChat Pay or crypto in USD, or UPI in INR (official payment methods read 2026-11-03). The terms (last updated October 20, 2026; read 2026-10-23) also allow one account per individual and reserve a permanent ban for multi-account use. The Go plan moves from a flat $10 wallet to per-model monthly credits (one shared $10 pool with larger allowances on selected models) at each subscriber's next billing cycle; the pricing-limits and usage-limits tables now give Go windows of $2 per 5 hours and $5 per week while the same page's migration FAQ says they stay at $3 and $6, its request estimate fell from ~15K to ~9K, and DeepSeek V4.1 Flash is boosted to the full $10 on Go (commandcode.ai/docs/resources/pricing-limits, read 2026-09-29). Mistral Large 4 is now listed across plans, including Go, with temporary boosted per-model monthly limits of $10/$40/$50 on Go/GOAT/Pro. These promotional boosts do not resolve the existing terms-versus-allowance conflict. Cloud agent access is available on every plan except Go, billed per second at $0.10/hour; it is not an included runtime allowance (pricing, read 2026-10-08). Claude Haiku 5.5 is now also available on GOAT at the $20 default per-model allowance (pricing-page release notes, read 2026-10-11); Pro and above retain access. The per-model allowance does not override the conflicting plan-wide terms. Glyph Cluster is a temporary free stealth preview limited to 1,000 requests/day (100 before October 31) with at least $1 credit balance; it is not a recurring allowance. The GLM-5.3 Flash 100-credit promotion ended October 17; GLM-5.3 Flash now has boosted per-model allowances of $10/$80/$90 on Go/GOAT/Pro "while capacity lasts" at unchanged token rates (official pricing-limits and provider docs, October 18). These model-specific figures do not resolve the separate contractual 1:1 ceiling, and do not increase our capacity assessment (October 15). Ling 3.1 Flash now has boosted per-model credits of $100 on GOAT and $110 on Pro at unchanged token rates. The temporary preview does not establish recurring included capacity. The DeepSeek V4.1 Flash temporary 100-credit promotion ended October 16; the October 17 per-model table showed $10/$80/$90 on Go/GOAT/Pro; the October 20 limits page no longer establishes these as a permanent boost. Beam 1.0 is a temporary free route while capacity lasts, with a $1 balance requirement and possible training on traffic; it is not a recurring subscription entitlement (official pricing and provider docs, October 17). The October 20 pricing page advertises 80 GOAT credits while its comparison table still says 70, and the terms retain the 1:1 API-cost ceiling. That unresolved conflict prevents a capacity figure. MiniMax M3.1 Flash Free is a temporary ZDR route with no daily cap and a $1 balance requirement, not a recurring token grant. The current provider table no longer publishes per-model ZDR check marks; historical symbols do not establish current retention terms. Ornith 1.5 122B Free is temporary through October 21 at 11 PM Pacific and requires a $1 balance. The DeepSeek October17 launch boost expired October21 at 01:00 UTC. Temporary rates do not establish a recurring allowance or resolve the GOAT 80-versus-70-credit and contractual 1:1-ceiling conflict (official provider, limits and terms, October 21). The October 22 usage page removes the expired Mistral promotion and shows a new temporary DeepSeek boost; neither adds recurring capacity. The GOAT price/credit and contractual ceiling discrepancies remain unresolved. The October 23 pages remove the MiMo V2.5/V2.5 Pro offers and expired MiniMax M3.1 Flash/Ornith free promotions. Paid MiniMax and Ornith routes remain, but no recurring capacity figure is established while the contractual and advertised credit totals conflict. October24: expired free TypeSafe/Ornith promotions are absent; GOAT100 listsLing3.1Flash, while the Pro300Qwen3.7Max offer is removed. The advertised balances still conflict with the contractual1:1 ceiling; no graded recurring allowance is assigned. October 27: the limits tables now advertise up to $100 GOAT and $110 Pro, with up to $20/$50 and $22/$55 five-hour/weekly limits respectively. These maxima correspond to Ling 3.1 Flash; GLM-5.2 remains the $70/$80 example. The October 28 pricing card now says up to $100 GOAT, and its comparison table agrees on up to $100 GOAT/$110 Pro. Mandatory processing charges and the contractual 1:1 API-cost ceiling remain unresolved, so no recurring token value is assigned. GPT-6.1 Terra is now listed by the Provider rate card; per-plan availability is recorded only where explicitly verified. October 28: current terms are dated October 21 and remove the assurance that CommandCode does not train its own models on submitted material. Pricing and provider pages remove the matching no-training statement; zero-retention enforcement remains conditional on provider support. These are revised disclosures, not evidence that training occurred. Mandatory fees and the contractual 1:1 API-cost ceiling still prevent a verified recurring token figure. October 31: DeepSeek V4.1 Flash and GLM-5.3 Flash boosted per-model allowances are $10/$100/$110 on Go/GOAT/Pro (were $10/$80/$90); the DeepSeek deal no longer says it has no end date. The Ling 3.1 Flash 1019 free promotion ended November 3. Crouton Alpha is now a temporary stealth preview, capped at 1,000 requests a day while it lasts, with a $1 balance requirement and possible provider retention of prompts and outputs. It is not a recurring allowance (official pricing and limits, November 4). Fees and the contractual 1:1 ceiling still prevent a figure.Conditions apply · checked 2026-10-31 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
CommandCode / Max 10xUSD 100/monthNot verifiedIndividual · base + unspecified mandatory fee$150 monthly usage reference, split $100 premium / $150 standard; $45 per 5-hour window; $90 per weekly window; per-model caps also apply. Reaching a shorter window can pause included usage before monthly credits run out. Optional paid credits can continue work. Mandatory processing fee is not included in the listed price. Advertised credit multiples are not guaranteed by the terms; additional credits depend on availability. Advertised monthly credits have five-hour, weekly and per-model limits. On the Max plans premium and standard models each carry their own limit ($100 premium / $150 standard on Max 10x, $200 / $300 on Max 20x), so the headline credit figure is not one spendable pool. The Go plan is the only plan without API access, and some models are geo-restricted by their providers (GPT-5.6 Luna and Gemini 3.7 Flash are unavailable in China). Per-model tier gating applies: several models are documented as available on GOAT and above only. The terms, reissued October 20, 2026 (read October 23), now state the ceiling outright: subscriptions are sold on the basis of underlying API cost and a buyer "may expect, at maximum, a one-to-one (1:1) ratio between the subscription amount paid and the API credits made available", with no guarantee of any greater ratio and actual usage possibly lower after model pricing, provider rate changes, taxes and overhead. Every advertised allowance above 1x the price - Go at $1 against a $10 usage reference, GOAT 70 credits for $10, Pro 80 for $20, both Max tiers at 1.5x - is therefore a marketing figure the terms disclaim. Additional, bonus, promotional and top-up credits are "while supplies last" and revocable at any time, and unused credits and capacity do not roll over to a later period. Coding-plan processing fees remain unspecified. Provider documentation gives a $1.01 processing fee (called a card fee until 2026-10-01) but conflicting recurring versus pay-as-you-go wording; plans and extra credits can be paid with a card, Alipay, WeChat Pay or crypto in USD, or UPI in INR (official payment methods read 2026-11-03). The terms (last updated October 20, 2026; read 2026-10-23) also allow one account per individual and reserve a permanent ban for multi-account use. The Go plan moves from a flat $10 wallet to per-model monthly credits (one shared $10 pool with larger allowances on selected models) at each subscriber's next billing cycle; the pricing-limits and usage-limits tables now give Go windows of $2 per 5 hours and $5 per week while the same page's migration FAQ says they stay at $3 and $6, its request estimate fell from ~15K to ~9K, and DeepSeek V4.1 Flash is boosted to the full $10 on Go (commandcode.ai/docs/resources/pricing-limits, read 2026-09-29). Mistral Large 4 is now listed across plans, including Go, with temporary boosted per-model monthly limits of $10/$40/$50 on Go/GOAT/Pro. These promotional boosts do not resolve the existing terms-versus-allowance conflict. Cloud agent access is available on every plan except Go, billed per second at $0.10/hour; it is not an included runtime allowance (pricing, read 2026-10-08). Claude Haiku 5.5 is now also available on GOAT at the $20 default per-model allowance (pricing-page release notes, read 2026-10-11); Pro and above retain access. The per-model allowance does not override the conflicting plan-wide terms. Glyph Cluster is a temporary free stealth preview limited to 1,000 requests/day (100 before October 31) with at least $1 credit balance; it is not a recurring allowance. The GLM-5.3 Flash 100-credit promotion ended October 17; GLM-5.3 Flash now has boosted per-model allowances of $10/$80/$90 on Go/GOAT/Pro "while capacity lasts" at unchanged token rates (official pricing-limits and provider docs, October 18). These model-specific figures do not resolve the separate contractual 1:1 ceiling, and do not increase our capacity assessment (October 15). Ling 3.1 Flash now has boosted per-model credits of $100 on GOAT and $110 on Pro at unchanged token rates. The temporary preview does not establish recurring included capacity. The DeepSeek V4.1 Flash temporary 100-credit promotion ended October 16; the October 17 per-model table showed $10/$80/$90 on Go/GOAT/Pro; the October 20 limits page no longer establishes these as a permanent boost. Beam 1.0 is a temporary free route while capacity lasts, with a $1 balance requirement and possible training on traffic; it is not a recurring subscription entitlement (official pricing and provider docs, October 17). The October 20 pricing page advertises 80 GOAT credits while its comparison table still says 70, and the terms retain the 1:1 API-cost ceiling. That unresolved conflict prevents a capacity figure. MiniMax M3.1 Flash Free is a temporary ZDR route with no daily cap and a $1 balance requirement, not a recurring token grant. The current provider table no longer publishes per-model ZDR check marks; historical symbols do not establish current retention terms. Ornith 1.5 122B Free is temporary through October 21 at 11 PM Pacific and requires a $1 balance. The DeepSeek October17 launch boost expired October21 at 01:00 UTC. Temporary rates do not establish a recurring allowance or resolve the GOAT 80-versus-70-credit and contractual 1:1-ceiling conflict (official provider, limits and terms, October 21). The October 22 usage page removes the expired Mistral promotion and shows a new temporary DeepSeek boost; neither adds recurring capacity. The GOAT price/credit and contractual ceiling discrepancies remain unresolved. The October 23 pages remove the MiMo V2.5/V2.5 Pro offers and expired MiniMax M3.1 Flash/Ornith free promotions. Paid MiniMax and Ornith routes remain, but no recurring capacity figure is established while the contractual and advertised credit totals conflict. October24: expired free TypeSafe/Ornith promotions are absent; GOAT100 listsLing3.1Flash, while the Pro300Qwen3.7Max offer is removed. The advertised balances still conflict with the contractual1:1 ceiling; no graded recurring allowance is assigned. October 27: the limits tables now advertise up to $100 GOAT and $110 Pro, with up to $20/$50 and $22/$55 five-hour/weekly limits respectively. These maxima correspond to Ling 3.1 Flash; GLM-5.2 remains the $70/$80 example. The October 28 pricing card now says up to $100 GOAT, and its comparison table agrees on up to $100 GOAT/$110 Pro. Mandatory processing charges and the contractual 1:1 API-cost ceiling remain unresolved, so no recurring token value is assigned. GPT-6.1 Terra is now listed by the Provider rate card; per-plan availability is recorded only where explicitly verified. October 28: current terms are dated October 21 and remove the assurance that CommandCode does not train its own models on submitted material. Pricing and provider pages remove the matching no-training statement; zero-retention enforcement remains conditional on provider support. These are revised disclosures, not evidence that training occurred. Mandatory fees and the contractual 1:1 API-cost ceiling still prevent a verified recurring token figure. October 31: DeepSeek V4.1 Flash and GLM-5.3 Flash boosted per-model allowances are $10/$100/$110 on Go/GOAT/Pro (were $10/$80/$90); the DeepSeek deal no longer says it has no end date. The Ling 3.1 Flash 1019 free promotion ended November 3. Crouton Alpha is now a temporary stealth preview, capped at 1,000 requests a day while it lasts, with a $1 balance requirement and possible provider retention of prompts and outputs. It is not a recurring allowance (official pricing and limits, November 4). Fees and the contractual 1:1 ceiling still prevent a figure.Conditions apply · checked 2026-10-31 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
CommandCode / Max 20xUSD 200/monthNot verifiedIndividual · base + unspecified mandatory fee$300 monthly usage reference, split $200 premium / $300 standard; $90 per 5-hour window; $180 per weekly window; per-model caps also apply. Reaching a shorter window can pause included usage before monthly credits run out. Optional paid credits can continue work. Mandatory processing fee is not included in the listed price. Advertised credit multiples are not guaranteed by the terms; additional credits depend on availability. Advertised monthly credits have five-hour, weekly and per-model limits. On the Max plans premium and standard models each carry their own limit ($100 premium / $150 standard on Max 10x, $200 / $300 on Max 20x), so the headline credit figure is not one spendable pool. The Go plan is the only plan without API access, and some models are geo-restricted by their providers (GPT-5.6 Luna and Gemini 3.7 Flash are unavailable in China). Per-model tier gating applies: several models are documented as available on GOAT and above only. The terms, reissued October 20, 2026 (read October 23), now state the ceiling outright: subscriptions are sold on the basis of underlying API cost and a buyer "may expect, at maximum, a one-to-one (1:1) ratio between the subscription amount paid and the API credits made available", with no guarantee of any greater ratio and actual usage possibly lower after model pricing, provider rate changes, taxes and overhead. Every advertised allowance above 1x the price - Go at $1 against a $10 usage reference, GOAT 70 credits for $10, Pro 80 for $20, both Max tiers at 1.5x - is therefore a marketing figure the terms disclaim. Additional, bonus, promotional and top-up credits are "while supplies last" and revocable at any time, and unused credits and capacity do not roll over to a later period. Coding-plan processing fees remain unspecified. Provider documentation gives a $1.01 processing fee (called a card fee until 2026-10-01) but conflicting recurring versus pay-as-you-go wording; plans and extra credits can be paid with a card, Alipay, WeChat Pay or crypto in USD, or UPI in INR (official payment methods read 2026-11-03). The terms (last updated October 20, 2026; read 2026-10-23) also allow one account per individual and reserve a permanent ban for multi-account use. The Go plan moves from a flat $10 wallet to per-model monthly credits (one shared $10 pool with larger allowances on selected models) at each subscriber's next billing cycle; the pricing-limits and usage-limits tables now give Go windows of $2 per 5 hours and $5 per week while the same page's migration FAQ says they stay at $3 and $6, its request estimate fell from ~15K to ~9K, and DeepSeek V4.1 Flash is boosted to the full $10 on Go (commandcode.ai/docs/resources/pricing-limits, read 2026-09-29). Mistral Large 4 is now listed across plans, including Go, with temporary boosted per-model monthly limits of $10/$40/$50 on Go/GOAT/Pro. These promotional boosts do not resolve the existing terms-versus-allowance conflict. Cloud agent access is available on every plan except Go, billed per second at $0.10/hour; it is not an included runtime allowance (pricing, read 2026-10-08). Claude Haiku 5.5 is now also available on GOAT at the $20 default per-model allowance (pricing-page release notes, read 2026-10-11); Pro and above retain access. The per-model allowance does not override the conflicting plan-wide terms. Glyph Cluster is a temporary free stealth preview limited to 1,000 requests/day (100 before October 31) with at least $1 credit balance; it is not a recurring allowance. The GLM-5.3 Flash 100-credit promotion ended October 17; GLM-5.3 Flash now has boosted per-model allowances of $10/$80/$90 on Go/GOAT/Pro "while capacity lasts" at unchanged token rates (official pricing-limits and provider docs, October 18). These model-specific figures do not resolve the separate contractual 1:1 ceiling, and do not increase our capacity assessment (October 15). Ling 3.1 Flash now has boosted per-model credits of $100 on GOAT and $110 on Pro at unchanged token rates. The temporary preview does not establish recurring included capacity. The DeepSeek V4.1 Flash temporary 100-credit promotion ended October 16; the October 17 per-model table showed $10/$80/$90 on Go/GOAT/Pro; the October 20 limits page no longer establishes these as a permanent boost. Beam 1.0 is a temporary free route while capacity lasts, with a $1 balance requirement and possible training on traffic; it is not a recurring subscription entitlement (official pricing and provider docs, October 17). The October 20 pricing page advertises 80 GOAT credits while its comparison table still says 70, and the terms retain the 1:1 API-cost ceiling. That unresolved conflict prevents a capacity figure. MiniMax M3.1 Flash Free is a temporary ZDR route with no daily cap and a $1 balance requirement, not a recurring token grant. The current provider table no longer publishes per-model ZDR check marks; historical symbols do not establish current retention terms. Ornith 1.5 122B Free is temporary through October 21 at 11 PM Pacific and requires a $1 balance. The DeepSeek October17 launch boost expired October21 at 01:00 UTC. Temporary rates do not establish a recurring allowance or resolve the GOAT 80-versus-70-credit and contractual 1:1-ceiling conflict (official provider, limits and terms, October 21). The October 22 usage page removes the expired Mistral promotion and shows a new temporary DeepSeek boost; neither adds recurring capacity. The GOAT price/credit and contractual ceiling discrepancies remain unresolved. The October 23 pages remove the MiMo V2.5/V2.5 Pro offers and expired MiniMax M3.1 Flash/Ornith free promotions. Paid MiniMax and Ornith routes remain, but no recurring capacity figure is established while the contractual and advertised credit totals conflict. October24: expired free TypeSafe/Ornith promotions are absent; GOAT100 listsLing3.1Flash, while the Pro300Qwen3.7Max offer is removed. The advertised balances still conflict with the contractual1:1 ceiling; no graded recurring allowance is assigned. October 27: the limits tables now advertise up to $100 GOAT and $110 Pro, with up to $20/$50 and $22/$55 five-hour/weekly limits respectively. These maxima correspond to Ling 3.1 Flash; GLM-5.2 remains the $70/$80 example. The October 28 pricing card now says up to $100 GOAT, and its comparison table agrees on up to $100 GOAT/$110 Pro. Mandatory processing charges and the contractual 1:1 API-cost ceiling remain unresolved, so no recurring token value is assigned. GPT-6.1 Terra is now listed by the Provider rate card; per-plan availability is recorded only where explicitly verified. October 28: current terms are dated October 21 and remove the assurance that CommandCode does not train its own models on submitted material. Pricing and provider pages remove the matching no-training statement; zero-retention enforcement remains conditional on provider support. These are revised disclosures, not evidence that training occurred. Mandatory fees and the contractual 1:1 API-cost ceiling still prevent a verified recurring token figure. October 31: DeepSeek V4.1 Flash and GLM-5.3 Flash boosted per-model allowances are $10/$100/$110 on Go/GOAT/Pro (were $10/$80/$90); the DeepSeek deal no longer says it has no end date. The Ling 3.1 Flash 1019 free promotion ended November 3. Crouton Alpha is now a temporary stealth preview, capped at 1,000 requests a day while it lasts, with a $1 balance requirement and possible provider retention of prompts and outputs. It is not a recurring allowance (official pricing and limits, November 4). Fees and the contractual 1:1 ceiling still prevent a figure.Conditions apply · checked 2026-10-31 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
CommandCode / ProviderUSD 15/monthNot verifiedIndividual · base + fee + usagePay-as-you-go inference; no markup claim Provider docs quote $15 plus a $1.01 card fee and $15 starting API credit, but also say no monthly minimums. Recurring pricing versus initial credit purchase remains ambiguous; confirm billing terms before comparing monthly value. Advertised monthly credits have five-hour, weekly and per-model limits. On the Max plans premium and standard models each carry their own limit ($100 premium / $150 standard on Max 10x, $200 / $300 on Max 20x), so the headline credit figure is not one spendable pool. The Go plan is the only plan without API access, and some models are geo-restricted by their providers (GPT-5.6 Luna and Gemini 3.7 Flash are unavailable in China). Per-model tier gating applies: several models are documented as available on GOAT and above only. The terms, reissued October 20, 2026 (read October 23), now state the ceiling outright: subscriptions are sold on the basis of underlying API cost and a buyer "may expect, at maximum, a one-to-one (1:1) ratio between the subscription amount paid and the API credits made available", with no guarantee of any greater ratio and actual usage possibly lower after model pricing, provider rate changes, taxes and overhead. Every advertised allowance above 1x the price - Go at $1 against a $10 usage reference, GOAT 70 credits for $10, Pro 80 for $20, both Max tiers at 1.5x - is therefore a marketing figure the terms disclaim. Additional, bonus, promotional and top-up credits are "while supplies last" and revocable at any time, and unused credits and capacity do not roll over to a later period. Coding-plan processing fees remain unspecified. Provider documentation gives a $1.01 processing fee (called a card fee until 2026-10-01) but conflicting recurring versus pay-as-you-go wording; plans and extra credits can be paid with a card, Alipay, WeChat Pay or crypto in USD, or UPI in INR (official payment methods read 2026-11-03). The terms (last updated October 20, 2026; read 2026-10-23) also allow one account per individual and reserve a permanent ban for multi-account use. The Go plan moves from a flat $10 wallet to per-model monthly credits (one shared $10 pool with larger allowances on selected models) at each subscriber's next billing cycle; the pricing-limits and usage-limits tables now give Go windows of $2 per 5 hours and $5 per week while the same page's migration FAQ says they stay at $3 and $6, its request estimate fell from ~15K to ~9K, and DeepSeek V4.1 Flash is boosted to the full $10 on Go (commandcode.ai/docs/resources/pricing-limits, read 2026-09-29). Mistral Large 4 is now listed across plans, including Go, with temporary boosted per-model monthly limits of $10/$40/$50 on Go/GOAT/Pro. These promotional boosts do not resolve the existing terms-versus-allowance conflict. Cloud agent access is available on every plan except Go, billed per second at $0.10/hour; it is not an included runtime allowance (pricing, read 2026-10-08). Claude Haiku 5.5 is now also available on GOAT at the $20 default per-model allowance (pricing-page release notes, read 2026-10-11); Pro and above retain access. The per-model allowance does not override the conflicting plan-wide terms. Glyph Cluster is a temporary free stealth preview limited to 1,000 requests/day (100 before October 31) with at least $1 credit balance; it is not a recurring allowance. The GLM-5.3 Flash 100-credit promotion ended October 17; GLM-5.3 Flash now has boosted per-model allowances of $10/$80/$90 on Go/GOAT/Pro "while capacity lasts" at unchanged token rates (official pricing-limits and provider docs, October 18). These model-specific figures do not resolve the separate contractual 1:1 ceiling, and do not increase our capacity assessment (October 15). Ling 3.1 Flash now has boosted per-model credits of $100 on GOAT and $110 on Pro at unchanged token rates. The temporary preview does not establish recurring included capacity. The DeepSeek V4.1 Flash temporary 100-credit promotion ended October 16; the October 17 per-model table showed $10/$80/$90 on Go/GOAT/Pro; the October 20 limits page no longer establishes these as a permanent boost. Beam 1.0 is a temporary free route while capacity lasts, with a $1 balance requirement and possible training on traffic; it is not a recurring subscription entitlement (official pricing and provider docs, October 17). The October 20 pricing page advertises 80 GOAT credits while its comparison table still says 70, and the terms retain the 1:1 API-cost ceiling. That unresolved conflict prevents a capacity figure. MiniMax M3.1 Flash Free is a temporary ZDR route with no daily cap and a $1 balance requirement, not a recurring token grant. The current provider table no longer publishes per-model ZDR check marks; historical symbols do not establish current retention terms. Ornith 1.5 122B Free is temporary through October 21 at 11 PM Pacific and requires a $1 balance. The DeepSeek October17 launch boost expired October21 at 01:00 UTC. Temporary rates do not establish a recurring allowance or resolve the GOAT 80-versus-70-credit and contractual 1:1-ceiling conflict (official provider, limits and terms, October 21). The October 22 usage page removes the expired Mistral promotion and shows a new temporary DeepSeek boost; neither adds recurring capacity. The GOAT price/credit and contractual ceiling discrepancies remain unresolved. The October 23 pages remove the MiMo V2.5/V2.5 Pro offers and expired MiniMax M3.1 Flash/Ornith free promotions. Paid MiniMax and Ornith routes remain, but no recurring capacity figure is established while the contractual and advertised credit totals conflict. October24: expired free TypeSafe/Ornith promotions are absent; GOAT100 listsLing3.1Flash, while the Pro300Qwen3.7Max offer is removed. The advertised balances still conflict with the contractual1:1 ceiling; no graded recurring allowance is assigned. October 27: the limits tables now advertise up to $100 GOAT and $110 Pro, with up to $20/$50 and $22/$55 five-hour/weekly limits respectively. These maxima correspond to Ling 3.1 Flash; GLM-5.2 remains the $70/$80 example. The October 28 pricing card now says up to $100 GOAT, and its comparison table agrees on up to $100 GOAT/$110 Pro. Mandatory processing charges and the contractual 1:1 API-cost ceiling remain unresolved, so no recurring token value is assigned. GPT-6.1 Terra is now listed by the Provider rate card; per-plan availability is recorded only where explicitly verified. October 28: current terms are dated October 21 and remove the assurance that CommandCode does not train its own models on submitted material. Pricing and provider pages remove the matching no-training statement; zero-retention enforcement remains conditional on provider support. These are revised disclosures, not evidence that training occurred. Mandatory fees and the contractual 1:1 API-cost ceiling still prevent a verified recurring token figure. October 31: DeepSeek V4.1 Flash and GLM-5.3 Flash boosted per-model allowances are $10/$100/$110 on Go/GOAT/Pro (were $10/$80/$90); the DeepSeek deal no longer says it has no end date. The Ling 3.1 Flash 1019 free promotion ended November 3. Crouton Alpha is now a temporary stealth preview, capped at 1,000 requests a day while it lasts, with a $1 balance requirement and possible provider retention of prompts and outputs. It is not a recurring allowance (official pricing and limits, November 4). Fees and the contractual 1:1 ceiling still prevent a figure.Conditions apply · checked 2026-10-31 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
T3 Chat / Paid membershipPrice not verifiedNot verifiedIndividual · standardBase allowance refills every 4 hours; a separate Overage allowance refills on monthly renewal. Both buckets empty means waiting for a refill. Long threads, attachments, search and expensive models consume more. The FAQ’s 20/200-point example is illustrative, not a published plan quota. The current FAQ documents usage-based buckets rather than a fixed message allowance. Public price verification remains incomplete; older message-count claims are not used.Needs verification · checked 2026-09-10 Source 1 · Source 2 · Source 3
Venice / FreeFreeNot verifiedIndividual · standard10 text + 15 image prompts/day Monthly prices verified in browser. Unlimited text applies to Pro-class models, while frontier models consume the premium pool. Credit conversion is 100/$1; the $18 plan has only 100 premium credits/month. Plus/Max offer limited rollover. Marketing privacy/unlimited claims were not independently tested. Unused premium credits bank forward on the two larger plans - two months on Pro Plus and three on Max - so an under-used month is not lost outright (venice.ai/pricing, read 2026-09-20).Documented · checked 2026-10-31 Source 1
Venice / ProUSD 18/monthNot verifiedIndividual · standard100 premium credits/month; separate Pro-model access Monthly prices verified in browser. Unlimited text applies to Pro-class models, while frontier models consume the premium pool. Credit conversion is 100/$1; the $18 plan has only 100 premium credits/month. Plus/Max offer limited rollover. Marketing privacy/unlimited claims were not independently tested. Unused premium credits bank forward on the two larger plans - two months on Pro Plus and three on Max - so an under-used month is not lost outright (venice.ai/pricing, read 2026-09-20).Documented · checked 2026-11-02 Source 1
Venice / Pro PlusUSD 68/monthNot verifiedIndividual · standard7,500 premium credits/month; separate Pro-model access Monthly prices verified in browser. Unlimited text applies to Pro-class models, while frontier models consume the premium pool. Credit conversion is 100/$1; the $18 plan has only 100 premium credits/month. Plus/Max offer limited rollover. Marketing privacy/unlimited claims were not independently tested. Unused premium credits bank forward on the two larger plans - two months on Pro Plus and three on Max - so an under-used month is not lost outright (venice.ai/pricing, read 2026-09-20).Documented · checked 2026-11-02 Source 1
Venice / MaxUSD 200/monthNot verifiedIndividual · standard22,500 premium credits/month; separate Pro-model access Monthly prices verified in browser. Unlimited text applies to Pro-class models, while frontier models consume the premium pool. Credit conversion is 100/$1; the $18 plan has only 100 premium credits/month. Plus/Max offer limited rollover. Marketing privacy/unlimited claims were not independently tested. Unused premium credits bank forward on the two larger plans - two months on Pro Plus and three on Max - so an under-used month is not lost outright (venice.ai/pricing, read 2026-09-20).Documented · checked 2026-11-02 Source 1
Manus / FreeFreeNot verifiedIndividual · standard300 daily credits; 1,500 monthly consumption cap Daily credits limited to Lite agent. Official membership, credit help and rendered monthly pricing reverified 2026-10-26: Pro starts at $20/4,000 or $40/8,000 credits per month, with 20 concurrent tasks and 20 scheduled automations. Free has 300 daily refresh credits, capped at 1,500 consumed per month, one concurrent task and two automations. Daily credits work only with Manus 2.0 Lite; Pro includes Manus 2.0, Max and Lite agent modes. Tokens, active VM work and third-party APIs consume credits. Subscription credits expire each monthly cycle, including annual plans; add-on credits roll over only while a paid subscription stays active. The monthly pricing selector confirms the $20/4,000 and $40/8,000 entry offers. Manus agent-mode names do not identify underlying ranked models, and credit consumption has no published model-token conversion, so these offers have no numeric value figure.Conditions apply · checked 2026-11-02 Source 1 · Source 2 · Source 3
Manus / Pro 4000 entryUSD 20/monthNot verifiedIndividual · starting; help-center priceStarts at 4,000 credits/month Checkout is dynamic; the official membership help verifies the entry price and monthly credit grant. Official membership, credit help and rendered monthly pricing reverified 2026-10-26: Pro starts at $20/4,000 or $40/8,000 credits per month, with 20 concurrent tasks and 20 scheduled automations. Free has 300 daily refresh credits, capped at 1,500 consumed per month, one concurrent task and two automations. Daily credits work only with Manus 2.0 Lite; Pro includes Manus 2.0, Max and Lite agent modes. Tokens, active VM work and third-party APIs consume credits. Subscription credits expire each monthly cycle, including annual plans; add-on credits roll over only while a paid subscription stays active. The monthly pricing selector confirms the $20/4,000 and $40/8,000 entry offers. Manus agent-mode names do not identify underlying ranked models, and credit consumption has no published model-token conversion, so these offers have no numeric value figure.Conditions apply · checked 2026-11-02 Source 1 · Source 2 · Source 3
Manus / Pro 8000 entryUSD 40/monthNot verifiedIndividual · starting; help-center priceStarts at 8,000 credits/month Checkout is dynamic; the official membership help verifies the entry price and monthly credit grant. Official membership, credit help and rendered monthly pricing reverified 2026-10-26: Pro starts at $20/4,000 or $40/8,000 credits per month, with 20 concurrent tasks and 20 scheduled automations. Free has 300 daily refresh credits, capped at 1,500 consumed per month, one concurrent task and two automations. Daily credits work only with Manus 2.0 Lite; Pro includes Manus 2.0, Max and Lite agent modes. Tokens, active VM work and third-party APIs consume credits. Subscription credits expire each monthly cycle, including annual plans; add-on credits roll over only while a paid subscription stays active. The monthly pricing selector confirms the $20/4,000 and $40/8,000 entry offers. Manus agent-mode names do not identify underlying ranked models, and credit consumption has no published model-token conversion, so these offers have no numeric value figure.Conditions apply · checked 2026-11-02 Source 1 · Source 2 · Source 3
Genspark / FreeFreeNot verifiedIndividual · standard100 credits/day until an unpublished lifetime allowance is used; the daily refill then stops. Official membership page publishes monthly and annual prices. Credits reset monthly even on annual billing. From September 18, 2026 AI Chat and AI Image include core models at no credit cost only up to a limit and then draw on monthly credits, while flagship models always use credits; Plus and Pro also receive Free Quota, a weekly balance for Standard-mode agent tasks whose size is not published. Subscriptions started before September 18, 2026 keep the previous unlimited chat and image experience through December 31, 2026. Cloud Computer requires a separate subscription. The membership guide now spells out commercial use on paid plans, including client work, while requiring clear source/context acknowledgement when republishing and compliance with applicable third-party rights. This does not grant rights the service does not own (membership guide, read 2026-10-07). AI Chat also has a separate per-user five-hour fair-use cap: buying credits does not lift it, and other agents keep working while chat waits. The live model picker is the source of truth; the public model examples are not exhaustive (AI Chat help, read 2026-10-11). Buying Team keeps the existing Plus or Pro subscription running until separately canceled; Team is a separate subscription, and canceling the individual plan does not cancel Team (official membership guide, November 4).Conditions apply · checked 2026-11-02 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
Genspark / Plus entryUSD 24.99/monthUSD 239.99Individual · standard10,000 credits/month; 50 GB storage; core-model AI Chat and AI Image free up to a limit, then credits; weekly Free Quota of unpublished size. Monthly credits expire at renewal. Agent retries and multi-model outputs consume credits. The unlimited chat and image benefit no longer applies to subscriptions started on or after September 18, 2026, so chat and image now compete with agent work for the same 10,000 credits. Plus and Pro include Nano Banana 2.1 in AI Image. AI Chat produces text only; image generation uses AI Image and its shared-credit rules. The membership page does not publish a per-image credit rate for this model (official membership benefits and AI Chat help, read October 25). Official membership page publishes monthly and annual prices. Credits reset monthly even on annual billing. From September 18, 2026 AI Chat and AI Image include core models at no credit cost only up to a limit and then draw on monthly credits, while flagship models always use credits; Plus and Pro also receive Free Quota, a weekly balance for Standard-mode agent tasks whose size is not published. Subscriptions started before September 18, 2026 keep the previous unlimited chat and image experience through December 31, 2026. Cloud Computer requires a separate subscription. The membership guide now spells out commercial use on paid plans, including client work, while requiring clear source/context acknowledgement when republishing and compliance with applicable third-party rights. This does not grant rights the service does not own (membership guide, read 2026-10-07). AI Chat also has a separate per-user five-hour fair-use cap: buying credits does not lift it, and other agents keep working while chat waits. The live model picker is the source of truth; the public model examples are not exhaustive (AI Chat help, read 2026-10-11). Buying Team keeps the existing Plus or Pro subscription running until separately canceled; Team is a separate subscription, and canceling the individual plan does not cancel Team (official membership guide, November 4).Documented · checked 2026-10-26 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
Genspark / Pro entryUSD 249.99/monthUSD 2,399.99Individual · standard125,000 credits/month; 1 TB storage; core-model AI Chat and AI Image free up to a higher limit, then credits; weekly Free Quota 10x the Plus balance. 12.5x the Plus credit allocation for about 10x the monthly price. Unused monthly credits expire; the higher tier pays off only if used. The unlimited chat and image benefit no longer applies to subscriptions started on or after September 18, 2026, but the 125,000-credit pool and a Free Quota ten times the Plus balance absorb that change. Plus and Pro include Nano Banana 2.1 in AI Image. AI Chat produces text only; image generation uses AI Image and its shared-credit rules. The membership page does not publish a per-image credit rate for this model (official membership benefits and AI Chat help, read October 25). Official membership page publishes monthly and annual prices. Credits reset monthly even on annual billing. From September 18, 2026 AI Chat and AI Image include core models at no credit cost only up to a limit and then draw on monthly credits, while flagship models always use credits; Plus and Pro also receive Free Quota, a weekly balance for Standard-mode agent tasks whose size is not published. Subscriptions started before September 18, 2026 keep the previous unlimited chat and image experience through December 31, 2026. Cloud Computer requires a separate subscription. The membership guide now spells out commercial use on paid plans, including client work, while requiring clear source/context acknowledgement when republishing and compliance with applicable third-party rights. This does not grant rights the service does not own (membership guide, read 2026-10-07). AI Chat also has a separate per-user five-hour fair-use cap: buying credits does not lift it, and other agents keep working while chat waits. The live model picker is the source of truth; the public model examples are not exhaustive (AI Chat help, read 2026-10-11). Buying Team keeps the existing Plus or Pro subscription running until separately canceled; Team is a separate subscription, and canceling the individual plan does not cancel Team (official membership guide, November 4).Documented · checked 2026-10-26 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
Notion AI / FreeFree trialNot verifiedWorkspace · standardTrial AI only USD currency and monthly billing verified. Free and Plus only trial AI; Business includes recurring AI functionality. Custom autonomous agents consume separately purchased Notion credits. An existing workplace license may already cover the person; paid seats are not household seats. Notion Agent (chat, generate, autofill, translate) is a limited trial on Free and Plus and included on Business; Custom Agents are free to try and then cost $10 per 1,000 monthly Notion credits, and Workers (beta custom code) require Notion credits (notion.com/pricing, read 2026-10-19; the earlier 15 October 2026 start date for custom code extensions is no longer printed there).Trial only · checked 2026-11-02 Source 1
Notion AI / PlusUSD 12/monthUSD 120Workspace · per memberTrial AI only; not an ongoing AI entitlement Annual total derived from $10/member/month billed yearly. USD currency and monthly billing verified. Free and Plus only trial AI; Business includes recurring AI functionality. Custom autonomous agents consume separately purchased Notion credits. An existing workplace license may already cover the person; paid seats are not household seats. Notion Agent (chat, generate, autofill, translate) is a limited trial on Free and Plus and included on Business; Custom Agents are free to try and then cost $10 per 1,000 monthly Notion credits, and Workers (beta custom code) require Notion credits (notion.com/pricing, read 2026-10-19; the earlier 15 October 2026 start date for custom code extensions is no longer printed there).Documented · checked 2026-11-02 Source 1
Notion AI / BusinessUSD 24/monthUSD 240Workspace · per memberNotion Agent, meeting notes and search; custom agents extra Annual total derived from $20/member/month billed yearly. USD currency and monthly billing verified. Free and Plus only trial AI; Business includes recurring AI functionality. Custom autonomous agents consume separately purchased Notion credits. An existing workplace license may already cover the person; paid seats are not household seats. Notion Agent (chat, generate, autofill, translate) is a limited trial on Free and Plus and included on Business; Custom Agents are free to try and then cost $10 per 1,000 monthly Notion credits, and Workers (beta custom code) require Notion credits (notion.com/pricing, read 2026-10-19; the earlier 15 October 2026 start date for custom code extensions is no longer printed there).Documented · checked 2026-11-02 Source 1
BytePlus ModelArk Coding Plan / LiteUSD 10/monthNot verifiedIndividual · standardApprox. 1,900 requests/5h; 12,000/week; 24,000/month Permitted coding clients only; model-dependent quota consumption. Browser rendering verified September pricing and the current subscription overview. The allowance is shared across permitted coding tools, with five-hour/weekly/monthly controls. Model invocations consume quota, and model coefficients affect effective capacity. General-purpose API usage is outside this subscription. Unused quota does not turn into cash or an unrestricted API balance. The two watched ModelArk docs establish the allowance and the plan rules but not the price: 1544106 is the API token-rate table and 1925114's subscription overview prints no figure, so the $10/$50 recurring prices are re-verifiable only on the signed-in console purchase page (re-checked 2026-09-22; byteplus.com/product/modelark publishes per-token API rates only).Needs a fresh check · checked 2026-09-20 Source 1 · Source 2
BytePlus ModelArk Coding Plan / ProUSD 50/monthNot verifiedIndividual · standardApprox. 9,500 requests/5h; 60,000/week; 120,000/month Permitted coding clients only; model-dependent quota consumption. Browser rendering verified September pricing and the current subscription overview. The allowance is shared across permitted coding tools, with five-hour/weekly/monthly controls. Model invocations consume quota, and model coefficients affect effective capacity. General-purpose API usage is outside this subscription. Unused quota does not turn into cash or an unrestricted API balance. The two watched ModelArk docs establish the allowance and the plan rules but not the price: 1544106 is the API token-rate table and 1925114's subscription overview prints no figure, so the $10/$50 recurring prices are re-verifiable only on the signed-in console purchase page (re-checked 2026-09-22; byteplus.com/product/modelark publishes per-token API rates only).Needs a fresh check · checked 2026-09-20 Source 1 · Source 2
Google AI / Gemini / FreeFreeNot verifiedIndividual · standardVariable Gemini access; separate creative limits Gemini uses compute-based five-hour and weekly controls; Flow credits are separate. US pricing and country eligibility apply. Gemini 4 Argon is included in AI Pro and Ultra in the Gemini app and Search. Paid Pro, eligible premium-device Pro trials and Ultra include Argon in the Antigravity app, CLI, standalone IDE and IDE extensions (official model matrix, October 27). Paid non-trial Pro and Ultra include Claude Opus 5.5 and Sonnet 5.5 thinking in Antigravity. Free and Plus do not. Gemini 3.8 Flash is available across Antigravity tiers. Older Opus/Sonnet 4.6 and GPT-OSS access there ends November 18 (official model matrix reviewed 2026-11-03). Gemini CLI and the Code Assist IDE extensions stopped serving free individual, Google AI Pro and Ultra requests on June 18, 2026. Consumer coding now uses Antigravity CLI, app and IDE extensions. Its included quota is work-dependent, with five-hour and weekly controls for paid plans; the old Gemini CLI request counts do not apply. Enterprise Code Assist and paid API-key access are separate.Documented · checked 2026-10-31 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6
Google AI / Gemini / AI PlusUSD 4.99/monthNot verifiedIndividual · standard2x Free Gemini usage; 200 Flow credits/month Nano Banana 2.1 image generation and editing is included in the Gemini app on Google AI Plus, Pro and Ultra. Its image allowance is separate from text usage; Google does not publish a 2.1-specific image count on the access page (official US image-generation page, read October 25). Gemini uses compute-based five-hour and weekly controls; Flow credits are separate. US pricing and country eligibility apply. Gemini 4 Argon is included in AI Pro and Ultra in the Gemini app and Search. Paid Pro, eligible premium-device Pro trials and Ultra include Argon in the Antigravity app, CLI, standalone IDE and IDE extensions (official model matrix, October 27). Paid non-trial Pro and Ultra include Claude Opus 5.5 and Sonnet 5.5 thinking in Antigravity. Free and Plus do not. Gemini 3.8 Flash is available across Antigravity tiers. Older Opus/Sonnet 4.6 and GPT-OSS access there ends November 18 (official model matrix reviewed 2026-11-03). Gemini CLI and the Code Assist IDE extensions stopped serving free individual, Google AI Pro and Ultra requests on June 18, 2026. Consumer coding now uses Antigravity CLI, app and IDE extensions. Its included quota is work-dependent, with five-hour and weekly controls for paid plans; the old Gemini CLI request counts do not apply. Enterprise Code Assist and paid API-key access are separate.Documented · checked 2026-10-31 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
Google AI / Gemini / AI ProUSD 19.99/monthNot verifiedIndividual · standard4x Free Gemini usage; 1,000 Flow credits/month Nano Banana 2.1 image generation and editing is included in the Gemini app on Google AI Plus, Pro and Ultra. Its image allowance is separate from text usage; Google does not publish a 2.1-specific image count on the access page (official US image-generation page, read October 25). Gemini uses compute-based five-hour and weekly controls; Flow credits are separate. US pricing and country eligibility apply. Gemini 4 Argon is included in AI Pro and Ultra in the Gemini app and Search. Paid Pro, eligible premium-device Pro trials and Ultra include Argon in the Antigravity app, CLI, standalone IDE and IDE extensions (official model matrix, October 27). Paid non-trial Pro and Ultra include Claude Opus 5.5 and Sonnet 5.5 thinking in Antigravity. Free and Plus do not. Gemini 3.8 Flash is available across Antigravity tiers. Older Opus/Sonnet 4.6 and GPT-OSS access there ends November 18 (official model matrix reviewed 2026-11-03). Gemini CLI and the Code Assist IDE extensions stopped serving free individual, Google AI Pro and Ultra requests on June 18, 2026. Consumer coding now uses Antigravity CLI, app and IDE extensions. Its included quota is work-dependent, with five-hour and weekly controls for paid plans; the old Gemini CLI request counts do not apply. Enterprise Code Assist and paid API-key access are separate.Documented · checked 2026-11-01 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7
Google AI / Gemini / AI Ultra 5xUSD 99.99/monthNot verifiedIndividual · standard5x Pro Gemini usage; 10,000 Flow credits/month The 5x multiplier covers the Gemini app and Google Antigravity (blog.google, read 2026-09-16); Antigravity publishes no quota numbers. Jules retains its separate task allowance. Consumer Gemini CLI access ended June 18, 2026; Antigravity CLI uses the plan's shared work-dependent quota, with no published request count. Nano Banana 2.1 image generation and editing is included in the Gemini app on Google AI Plus, Pro and Ultra. Its image allowance is separate from text usage; Google does not publish a 2.1-specific image count on the access page (official US image-generation page, read October 25). Gemini uses compute-based five-hour and weekly controls; Flow credits are separate. US pricing and country eligibility apply. Gemini 4 Argon is included in AI Pro and Ultra in the Gemini app and Search. Paid Pro, eligible premium-device Pro trials and Ultra include Argon in the Antigravity app, CLI, standalone IDE and IDE extensions (official model matrix, October 27). Paid non-trial Pro and Ultra include Claude Opus 5.5 and Sonnet 5.5 thinking in Antigravity. Free and Plus do not. Gemini 3.8 Flash is available across Antigravity tiers. Older Opus/Sonnet 4.6 and GPT-OSS access there ends November 18 (official model matrix reviewed 2026-11-03). Gemini CLI and the Code Assist IDE extensions stopped serving free individual, Google AI Pro and Ultra requests on June 18, 2026. Consumer coding now uses Antigravity CLI, app and IDE extensions. Its included quota is work-dependent, with five-hour and weekly controls for paid plans; the old Gemini CLI request counts do not apply. Enterprise Code Assist and paid API-key access are separate.Documented · checked 2026-10-31 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
Google AI / Gemini / AI Ultra 20xUSD 199.99/monthNot verifiedIndividual · standard20x Pro Gemini usage; 25,000 Flow credits/month The 20x multiplier covers the Gemini app and Google Antigravity (blog.google, read 2026-09-16); Antigravity publishes no quota numbers. Jules retains its separate task allowance. Consumer Gemini CLI access ended June 18, 2026; Antigravity CLI uses the plan's shared work-dependent quota, with no published request count. Nano Banana 2.1 image generation and editing is included in the Gemini app on Google AI Plus, Pro and Ultra. Its image allowance is separate from text usage; Google does not publish a 2.1-specific image count on the access page (official US image-generation page, read October 25). Gemini uses compute-based five-hour and weekly controls; Flow credits are separate. US pricing and country eligibility apply. Gemini 4 Argon is included in AI Pro and Ultra in the Gemini app and Search. Paid Pro, eligible premium-device Pro trials and Ultra include Argon in the Antigravity app, CLI, standalone IDE and IDE extensions (official model matrix, October 27). Paid non-trial Pro and Ultra include Claude Opus 5.5 and Sonnet 5.5 thinking in Antigravity. Free and Plus do not. Gemini 3.8 Flash is available across Antigravity tiers. Older Opus/Sonnet 4.6 and GPT-OSS access there ends November 18 (official model matrix reviewed 2026-11-03). Gemini CLI and the Code Assist IDE extensions stopped serving free individual, Google AI Pro and Ultra requests on June 18, 2026. Consumer coding now uses Antigravity CLI, app and IDE extensions. Its included quota is work-dependent, with five-hour and weekly controls for paid plans; the old Gemini CLI request counts do not apply. Enterprise Code Assist and paid API-key access are separate.Documented · checked 2026-10-31 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
MiniMax M Plan / GoUSD 22/monthUSD 220Individual · standard1× Go shared usage; non-video five-hour + weekly limits Includes M3.1 Flash text, image, audio, web search and multimodal understanding. No H3 video. Text, image, audio and supported video share one quota across tools. Non-video use has five-hour and weekly caps; video only the weekly cap. Absolute capacity is undisclosed, video consumes more, and unused subscription allowance does not roll over. Credit packs cost 1,000/$1 and expire after 365 days. Annual billing pays for ten months. Token Plan is closed to new purchases; existing auto-renewing subscribers retain benefits, but cannot switch old tiers or re-enable renewal after disabling it. Upgrading to M Plan is irreversible and ends historical Token Plan bonuses. M3.1 Flash replaces the preview in all three current M Plan tiers (official FAQ, October 20); this does not establish a new absolute allowance.Documented · checked 2026-11-03 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7 · Source 8
MiniMax M Plan / ExploreUSD 55/monthUSD 550Individual · standard3× Go shared usage; non-video five-hour + weekly limits Includes M3.1 Flash text, image, audio, web search and multimodal understanding. Includes H3 video against the shared weekly allowance. Text, image, audio and supported video share one quota across tools. Non-video use has five-hour and weekly caps; video only the weekly cap. Absolute capacity is undisclosed, video consumes more, and unused subscription allowance does not roll over. Credit packs cost 1,000/$1 and expire after 365 days. Annual billing pays for ten months. Token Plan is closed to new purchases; existing auto-renewing subscribers retain benefits, but cannot switch old tiers or re-enable renewal after disabling it. Upgrading to M Plan is irreversible and ends historical Token Plan bonuses. M3.1 Flash replaces the preview in all three current M Plan tiers (official FAQ, October 20); this does not establish a new absolute allowance.Documented · checked 2026-11-03 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7 · Source 8
MiniMax M Plan / BuildUSD 132/monthUSD 1,320Individual · standard7.5× Go shared usage; non-video five-hour + weekly limits Includes M3.1 Flash text, image, audio, web search and multimodal understanding. Includes H3 video against the shared weekly allowance. Text, image, audio and supported video share one quota across tools. Non-video use has five-hour and weekly caps; video only the weekly cap. Absolute capacity is undisclosed, video consumes more, and unused subscription allowance does not roll over. Credit packs cost 1,000/$1 and expire after 365 days. Annual billing pays for ten months. Token Plan is closed to new purchases; existing auto-renewing subscribers retain benefits, but cannot switch old tiers or re-enable renewal after disabling it. Upgrading to M Plan is irreversible and ends historical Token Plan bonuses. M3.1 Flash replaces the preview in all three current M Plan tiers (official FAQ, October 20); this does not establish a new absolute allowance.Documented · checked 2026-11-03 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7 · Source 8
Kimi membership / Kimi Code / PlusUSD 19/monthUSD 180Individual · standard1× Plus shared Agent quota; token-metered monthly pool; Code five-hour limit, no weekly limit for new plans USD monthly and annual purchase options verified in the browser on 2026-10-22. Absolute token quota is not published. Two parallel Agent tasks; K3 coding at 256K context. New plans and legacy subscriptions have different limits. All features share a token-metered monthly pool, with credits expiring each cycle. New Plus and above include Kimi Code and K3 with a rolling five-hour limit but no weekly limit; legacy subscriptions retain their weekly limit. K3 1M context and HighSpeed coding require new Pro or above, or legacy Allegretto or above. K3 1M coding consumes twice the quota of K3 256K. Extra Usage is separately billed. Plan eligibility follows the signed-in purchase page. Kimi Claw cloud hosts and published websites also consume the shared credits. International USD Plus/Pro/Max/Ultra pricing is published separately from the new mainland-China CNY Go/Plus/Pro/Max offers. Kimi Code has six selectable IDs: K3.1Pro and its Highspeed route, K3 and K3-256K, K2.8Preview and K2.7CodeHighspeed. An active Kimi Claw host consumes about 0.6% of monthly credits daily even while idle; a published Agent website consumes about 0.08% (official international pricing and model guide, reviewed 2026-10-14). Mainland-China Ultra is now advertised at CNY 1,399/month or CNY 13,428/year with 28× Go credits, four parallel tasks, 25 scheduled tasks/widgets, 100 projects and 50 GB storage. These regional terms do not replace the separately priced international USD Ultra offer (official regional and international pricing pages, October 20). October24 official model matrix: K3.1Pro requires newPlus+/legacyModerato+;1Mcontext requires newPro+/legacyAllegretto+. K3.1ProHighspeed requires newPro+/legacyAllegretto+, offers identical capabilities at2–3×speed and about2×quota usage. Thinking is mandatory. K3 remains selectable; the October5 measured K3 pools are retained at their original evidence date. No published conversion maps them to K3.1Pro.Conditions apply · checked 2026-11-02 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6
Kimi membership / Kimi Code / ProUSD 39/monthUSD 372Individual · standard2× Plus shared Agent quota; token-metered monthly pool; Code five-hour limit, no weekly limit for new plans USD monthly and annual purchase options verified in the browser on 2026-10-22. Absolute token quota is not published. K3 coding up to 1M context and HighSpeed available; 1M context spends twice the coding quota. New plans and legacy subscriptions have different limits. All features share a token-metered monthly pool, with credits expiring each cycle. New Plus and above include Kimi Code and K3 with a rolling five-hour limit but no weekly limit; legacy subscriptions retain their weekly limit. K3 1M context and HighSpeed coding require new Pro or above, or legacy Allegretto or above. K3 1M coding consumes twice the quota of K3 256K. Extra Usage is separately billed. Plan eligibility follows the signed-in purchase page. Kimi Claw cloud hosts and published websites also consume the shared credits. International USD Plus/Pro/Max/Ultra pricing is published separately from the new mainland-China CNY Go/Plus/Pro/Max offers. Kimi Code has six selectable IDs: K3.1Pro and its Highspeed route, K3 and K3-256K, K2.8Preview and K2.7CodeHighspeed. An active Kimi Claw host consumes about 0.6% of monthly credits daily even while idle; a published Agent website consumes about 0.08% (official international pricing and model guide, reviewed 2026-10-14). Mainland-China Ultra is now advertised at CNY 1,399/month or CNY 13,428/year with 28× Go credits, four parallel tasks, 25 scheduled tasks/widgets, 100 projects and 50 GB storage. These regional terms do not replace the separately priced international USD Ultra offer (official regional and international pricing pages, October 20). October24 official model matrix: K3.1Pro requires newPlus+/legacyModerato+;1Mcontext requires newPro+/legacyAllegretto+. K3.1ProHighspeed requires newPro+/legacyAllegretto+, offers identical capabilities at2–3×speed and about2×quota usage. Thinking is mandatory. K3 remains selectable; the October5 measured K3 pools are retained at their original evidence date. No published conversion maps them to K3.1Pro.Conditions apply · checked 2026-11-02 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6
Kimi membership / Kimi Code / MaxUSD 99/monthUSD 948Individual · standard5× Plus shared Agent quota; token-metered monthly pool; Code five-hour limit, no weekly limit for new plans USD monthly and annual purchase options verified in the browser on 2026-10-22. Absolute token quota is not published. K3 coding up to 1M context and HighSpeed available; 1M context spends twice the coding quota. New plans and legacy subscriptions have different limits. All features share a token-metered monthly pool, with credits expiring each cycle. New Plus and above include Kimi Code and K3 with a rolling five-hour limit but no weekly limit; legacy subscriptions retain their weekly limit. K3 1M context and HighSpeed coding require new Pro or above, or legacy Allegretto or above. K3 1M coding consumes twice the quota of K3 256K. Extra Usage is separately billed. Plan eligibility follows the signed-in purchase page. Kimi Claw cloud hosts and published websites also consume the shared credits. International USD Plus/Pro/Max/Ultra pricing is published separately from the new mainland-China CNY Go/Plus/Pro/Max offers. Kimi Code has six selectable IDs: K3.1Pro and its Highspeed route, K3 and K3-256K, K2.8Preview and K2.7CodeHighspeed. An active Kimi Claw host consumes about 0.6% of monthly credits daily even while idle; a published Agent website consumes about 0.08% (official international pricing and model guide, reviewed 2026-10-14). Mainland-China Ultra is now advertised at CNY 1,399/month or CNY 13,428/year with 28× Go credits, four parallel tasks, 25 scheduled tasks/widgets, 100 projects and 50 GB storage. These regional terms do not replace the separately priced international USD Ultra offer (official regional and international pricing pages, October 20). October24 official model matrix: K3.1Pro requires newPlus+/legacyModerato+;1Mcontext requires newPro+/legacyAllegretto+. K3.1ProHighspeed requires newPro+/legacyAllegretto+, offers identical capabilities at2–3×speed and about2×quota usage. Thinking is mandatory. K3 remains selectable; the October5 measured K3 pools are retained at their original evidence date. No published conversion maps them to K3.1Pro.Conditions apply · checked 2026-11-02 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6
Kimi membership / Kimi Code / UltraUSD 199/monthUSD 1,908Individual · standard10× Plus shared Agent quota; token-metered monthly pool; Code five-hour limit, no weekly limit for new plans USD monthly and annual purchase options verified in the browser on 2026-10-22. Absolute token quota is not published. K3 coding up to 1M context and HighSpeed available; 1M context spends twice the coding quota. The international page specifies four parallel tasks, 25 scheduled tasks/widgets, 100 projects and 50 GB storage; its Agent pool is 2× Max / 10× Plus (October 20). New plans and legacy subscriptions have different limits. All features share a token-metered monthly pool, with credits expiring each cycle. New Plus and above include Kimi Code and K3 with a rolling five-hour limit but no weekly limit; legacy subscriptions retain their weekly limit. K3 1M context and HighSpeed coding require new Pro or above, or legacy Allegretto or above. K3 1M coding consumes twice the quota of K3 256K. Extra Usage is separately billed. Plan eligibility follows the signed-in purchase page. Kimi Claw cloud hosts and published websites also consume the shared credits. International USD Plus/Pro/Max/Ultra pricing is published separately from the new mainland-China CNY Go/Plus/Pro/Max offers. Kimi Code has six selectable IDs: K3.1Pro and its Highspeed route, K3 and K3-256K, K2.8Preview and K2.7CodeHighspeed. An active Kimi Claw host consumes about 0.6% of monthly credits daily even while idle; a published Agent website consumes about 0.08% (official international pricing and model guide, reviewed 2026-10-14). Mainland-China Ultra is now advertised at CNY 1,399/month or CNY 13,428/year with 28× Go credits, four parallel tasks, 25 scheduled tasks/widgets, 100 projects and 50 GB storage. These regional terms do not replace the separately priced international USD Ultra offer (official regional and international pricing pages, October 20). October24 official model matrix: K3.1Pro requires newPlus+/legacyModerato+;1Mcontext requires newPro+/legacyAllegretto+. K3.1ProHighspeed requires newPro+/legacyAllegretto+, offers identical capabilities at2–3×speed and about2×quota usage. Thinking is mandatory. K3 remains selectable; the October5 measured K3 pools are retained at their original evidence date. No published conversion maps them to K3.1Pro.Conditions apply · checked 2026-11-02 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6
Cursor (legacy plans) / Cursor ProUSD 20/monthNot verifiedIndividual · standardCursor Models and Other Models pools; tab completions included. Two separate pools. Third-party usage consumes allowance at API rates. Grok Bot is included, with an unquantified weekly allowance (official FAQ, read 2026-10-09). Closed to new signups on October 12, 2026. Existing subscribers retain their plan, price and features unless they choose the permanent switch to The X Subscription. API requests are billed separately from the first request; the subscription does not include an API credit pool (official pricing, October 15).Retired · checked 2026-10-18 Source 1 · Source 2 · Source 3 · Source 4
Cursor (legacy plans) / Cursor Pro PlusUSD 60/monthNot verifiedIndividual · standardHigher agent allowance; pool amounts are not quantified here. Two separate pools. Third-party usage consumes allowance at API rates. Grok Bot is included on this tier (docs.x.ai/grok-bot/faq, read 2026-09-16); its weekly usage is not quantified. Closed to new signups on October 12, 2026. Existing subscribers retain their plan, price and features unless they choose the permanent switch to The X Subscription. API requests are billed separately from the first request; the subscription does not include an API credit pool (official pricing, October 15).Retired · checked 2026-10-18 Source 1 · Source 2 · Source 3 · Source 4
Cursor (legacy plans) / Cursor UltraUSD 200/monthNot verifiedIndividual · standardHigher agent allowance; pool amounts are not quantified here. Two separate pools. Third-party usage consumes allowance at API rates. Grok Bot is included on this tier (docs.x.ai/grok-bot/faq, read 2026-09-16); its weekly usage is not quantified. Closed to new signups on October 12, 2026. Existing subscribers retain their plan, price and features unless they choose the permanent switch to The X Subscription. API requests are billed separately from the first request; the subscription does not include an API credit pool (official pricing, October 15).Retired · checked 2026-10-18 Source 1 · Source 2 · Source 3 · Source 4
The X Subscription / PlusUSD 8/monthUSD 80Individual · standardBasic shared weekly Grok usage. Absolute task and token counts are not published. Grok models only; cloud agents, automations, Grok Bot and third-party model access require Premium or above. One weekly Grok usage pool is shared across Grok and Cursor. On Premium and above, third-party models in Cursor are billed separately from the first use; they are not part of the included allowance. Taxes and fees are extra. Switching from a legacy plan is optional and permanent. The October 28 Grok Bot FAQ confirms Bot is included on Premium, Super and Ultra, using the same weekly pool; Plus and Free exclude it. Bot selects models automatically without a model picker. For legacy accounts with linked Cursor and SuperGrok subscriptions, Bot uses the larger allowance rather than adding them, and both subscriptions continue billing separately.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7
The X Subscription / PremiumUSD 30/monthUSD 300Individual · standard4× Plus shared weekly Grok usage. Absolute task and token counts are not published. Includes the latest Grok models, Grok Bot, Cursor cloud agents, automations and code reviews; third-party usage costs extra. One weekly Grok usage pool is shared across Grok and Cursor. On Premium and above, third-party models in Cursor are billed separately from the first use; they are not part of the included allowance. Taxes and fees are extra. Switching from a legacy plan is optional and permanent. The October 28 Grok Bot FAQ confirms Bot is included on Premium, Super and Ultra, using the same weekly pool; Plus and Free exclude it. Bot selects models automatically without a model picker. For legacy accounts with linked Cursor and SuperGrok subscriptions, Bot uses the larger allowance rather than adding them, and both subscriptions continue billing separately.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7
The X Subscription / SuperUSD 100/monthUSD 1,000Individual · standard15× Plus shared weekly Grok usage. Absolute task and token counts are not published. Includes the latest Grok models, Grok Bot, Cursor cloud agents, automations and code reviews; third-party usage costs extra. One weekly Grok usage pool is shared across Grok and Cursor. On Premium and above, third-party models in Cursor are billed separately from the first use; they are not part of the included allowance. Taxes and fees are extra. Switching from a legacy plan is optional and permanent. The October 28 Grok Bot FAQ confirms Bot is included on Premium, Super and Ultra, using the same weekly pool; Plus and Free exclude it. Bot selects models automatically without a model picker. For legacy accounts with linked Cursor and SuperGrok subscriptions, Bot uses the larger allowance rather than adding them, and both subscriptions continue billing separately.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7
The X Subscription / UltraUSD 200/monthUSD 2,000Individual · standard50× Plus shared weekly Grok usage. Absolute task and token counts are not published. Includes the latest Grok models, Grok Bot, Cursor cloud agents, automations and code reviews; third-party usage costs extra. One weekly Grok usage pool is shared across Grok and Cursor. On Premium and above, third-party models in Cursor are billed separately from the first use; they are not part of the included allowance. Taxes and fees are extra. Switching from a legacy plan is optional and permanent. The October 28 Grok Bot FAQ confirms Bot is included on Premium, Super and Ultra, using the same weekly pool; Plus and Free exclude it. Bot selects models automatically without a model picker. For legacy accounts with linked Cursor and SuperGrok subscriptions, Bot uses the larger allowance rather than adding them, and both subscriptions continue billing separately.Documented · checked 2026-11-04 Source 1 · Source 2 · Source 3 · Source 4 · Source 5 · Source 6 · Source 7
Meta Muse Code / Everyday UsageUSD 5/monthNot verifiedIndividual · standard10–50 prompts per five hours, including image and video uploads Voice mode and web search included. Muse Wonder Full is not included. Subscription usage works only through the Muse Code CLI with the automatically connected subscription credential; additional API keys are pay-as-you-go. A Meta Account and age 18 or local age of majority are required, and benefits and availability vary by region. Contributor mode is on by default and permits training on prompts, code, files, context and responses in exchange for more usage. It can be turned off for future content. Muse Wonder is included on every plan; the distinct Muse Wonder Full model requires Power or Maximum. Selecting max reasoning effort on Wonder does not select Full. Both are available in Standard or Contributor mode on eligible plans. Usage resets in rolling five-hour windows; at the limit upgrade or wait, while Maximum users must wait. The published 10–50 prompt range varies with work and is not a guaranteed completed-task count. Monthly renewal is automatic; no annual price is documented. Muse Wonder and Muse Wonder Full are unavailable in South Korea on both Model API and Muse Code (official model matrix, October 20).Documented · checked 2026-11-01 Source 1 · Source 2 · Source 3
Meta Muse Code / High UsageUSD 15/monthNot verifiedIndividual · standard5x the Everyday Usage allowance, with rolling five-hour limits Voice mode and web search included. Muse Wonder Full is not included. Subscription usage works only through the Muse Code CLI with the automatically connected subscription credential; additional API keys are pay-as-you-go. A Meta Account and age 18 or local age of majority are required, and benefits and availability vary by region. Contributor mode is on by default and permits training on prompts, code, files, context and responses in exchange for more usage. It can be turned off for future content. Muse Wonder is included on every plan; the distinct Muse Wonder Full model requires Power or Maximum. Selecting max reasoning effort on Wonder does not select Full. Both are available in Standard or Contributor mode on eligible plans. Usage resets in rolling five-hour windows; at the limit upgrade or wait, while Maximum users must wait. The published 10–50 prompt range varies with work and is not a guaranteed completed-task count. Monthly renewal is automatic; no annual price is documented. Muse Wonder and Muse Wonder Full are unavailable in South Korea on both Model API and Muse Code (official model matrix, October 20).Documented · checked 2026-11-02 Source 1 · Source 2 · Source 3
Meta Muse Code / Power UsageUSD 50/monthNot verifiedIndividual · standard20x the Everyday Usage allowance, with rolling five-hour limits Voice mode and web search included. Power and Maximum include Muse Wonder Full. Subscription usage works only through the Muse Code CLI with the automatically connected subscription credential; additional API keys are pay-as-you-go. A Meta Account and age 18 or local age of majority are required, and benefits and availability vary by region. Contributor mode is on by default and permits training on prompts, code, files, context and responses in exchange for more usage. It can be turned off for future content. Muse Wonder is included on every plan; the distinct Muse Wonder Full model requires Power or Maximum. Selecting max reasoning effort on Wonder does not select Full. Both are available in Standard or Contributor mode on eligible plans. Usage resets in rolling five-hour windows; at the limit upgrade or wait, while Maximum users must wait. The published 10–50 prompt range varies with work and is not a guaranteed completed-task count. Monthly renewal is automatic; no annual price is documented. Muse Wonder and Muse Wonder Full are unavailable in South Korea on both Model API and Muse Code (official model matrix, October 20).Documented · checked 2026-11-01 Source 1 · Source 2 · Source 3
Meta Muse Code / Maximum UsageUSD 200/monthNot verifiedIndividual · standard80x the Everyday Usage allowance, with rolling five-hour limits Voice mode and web search included. Power and Maximum include Muse Wonder Full. Subscription usage works only through the Muse Code CLI with the automatically connected subscription credential; additional API keys are pay-as-you-go. A Meta Account and age 18 or local age of majority are required, and benefits and availability vary by region. Contributor mode is on by default and permits training on prompts, code, files, context and responses in exchange for more usage. It can be turned off for future content. Muse Wonder is included on every plan; the distinct Muse Wonder Full model requires Power or Maximum. Selecting max reasoning effort on Wonder does not select Full. Both are available in Standard or Contributor mode on eligible plans. Usage resets in rolling five-hour windows; at the limit upgrade or wait, while Maximum users must wait. The published 10–50 prompt range varies with work and is not a guaranteed completed-task count. Monthly renewal is automatic; no annual price is documented. Muse Wonder and Muse Wonder Full are unavailable in South Korea on both Model API and Muse Code (official model matrix, October 20).Documented · checked 2026-11-01 Source 1 · Source 2 · Source 3
Google AI / Gemini / Gemini Code Assist Free (retired)FreeNot verifiedIndividual · standardHistorical: 1,000 model requests/day with Google sign-in. Consumer service ended June 18, 2026. Gemini CLI and Code Assist IDE extensions stopped serving individual free accounts on June 18, 2026. Google directs consumers to Antigravity CLI and its work-dependent allowance. Enterprise licenses and paid API keys remain separate. Gemini uses compute-based five-hour and weekly controls; Flow credits are separate. US pricing and country eligibility apply. Gemini 4 Argon is included in AI Pro and Ultra in the Gemini app and Search. Paid Pro, eligible premium-device Pro trials and Ultra include Argon in the Antigravity app, CLI, standalone IDE and IDE extensions (official model matrix, October 27). Paid non-trial Pro and Ultra include Claude Opus 5.5 and Sonnet 5.5 thinking in Antigravity. Free and Plus do not. Gemini 3.8 Flash is available across Antigravity tiers. Older Opus/Sonnet 4.6 and GPT-OSS access there ends November 18 (official model matrix reviewed 2026-11-03). Gemini CLI and the Code Assist IDE extensions stopped serving free individual, Google AI Pro and Ultra requests on June 18, 2026. Consumer coding now uses Antigravity CLI, app and IDE extensions. Its included quota is work-dependent, with five-hour and weekly controls for paid plans; the old Gemini CLI request counts do not apply. Enterprise Code Assist and paid API-key access are separate.Retired · checked 2026-11-01 Source 1 · Source 2 · Source 3 · Source 4 · Source 5
ChatGPT / Codex / ChatGPT Pro 500USD 500/monthNot verifiedIndividual · standardThe highest included usage of the three Pro plans, with no five-hour limit; OpenAI doubled the Work and Codex allowance at the end of October; weekly limits can apply; no per-model estimates are published for Pro. Includes Ultrafast GPT-6 Astra and GPT-6.1 Sol and $500/month of eligible API Platform credit after activation. Billed monthly only. Ultrafast draws the included allowance at 8× the Standard rate and purchased credits at 6×. Paid Pro accounts active for at least seven days can link an API organization they own using the same email to claim monthly API credits. The first grant is prorated; unused credit expires each billing cycle. API organization/model eligibility and rate limits apply. Credits cover eligible API Platform usage, excluding Codex with API keys, and do not increase signed-in Work or Codex limits. Agent usage is shared. Task complexity and weekly limits affect capacity. GPT-5.5 retired from ChatGPT and Codex on October 14; Codex with ChatGPT stopped serving it at 19:00 UTC. API-key access is unaffected (official pricing and changelog, October 15).Documented · checked 2026-11-02 Source 1 · Source 2 · Source 3
ChatGPT / Codex / ChatGPT FreeFreeNot verifiedIndividual · standardLimited access for quick tasks. Agent usage is shared. Task complexity and weekly limits affect capacity. GPT-5.5 retired from ChatGPT and Codex on October 14; Codex with ChatGPT stopped serving it at 19:00 UTC. API-key access is unaffected (official pricing and changelog, October 15).Documented · checked 2026-11-02 Source 1
ChatGPT / Codex / ChatGPT GoUSD 8/monthNot verifiedIndividual · standardLightweight Codex use; no fixed allowance stated here. Agent usage is shared. Task complexity and weekly limits affect capacity. GPT-5.5 retired from ChatGPT and Codex on October 14; Codex with ChatGPT stopped serving it at 19:00 UTC. API-key access is unaffected (official pricing and changelog, October 15).Documented · checked 2026-11-02 Source 1
ChatGPT / Codex / ChatGPT PlusUSD 20/monthNot verifiedIndividual · standardEstimated local messages per 5 hours: GPT-6 Astra 5–45, GPT-6.1 Sol 15–160, GPT-6.1 Terra 35–285, GPT-6 Sol 15–150. Agent usage is shared. Task complexity and weekly limits affect capacity. GPT-5.5 retired from ChatGPT and Codex on October 14; Codex with ChatGPT stopped serving it at 19:00 UTC. API-key access is unaffected (official pricing and changelog, October 15).Documented · checked 2026-11-02 Source 1
ChatGPT / Codex / ChatGPT Pro 100USD 100/monthNot verifiedIndividual · standardMore usage than Plus with no five-hour limit; OpenAI doubled the Work and Codex allowance at the end of October; weekly limits can apply; no per-model estimates are published for Pro. Includes $100/month of eligible API Platform credit after activation. Agent usage is shared. Task complexity and weekly limits affect capacity. GPT-5.5 retired from ChatGPT and Codex on October 14; Codex with ChatGPT stopped serving it at 19:00 UTC. API-key access is unaffected (official pricing and changelog, October 15). Paid Pro accounts active for at least seven days can link an API organization they own using the same email to claim monthly API credits. The first grant is prorated; unused credit expires each billing cycle. API organization/model eligibility and rate limits apply. Credits cover eligible API Platform usage, excluding Codex with API keys, and do not increase signed-in Work or Codex limits.Documented · checked 2026-11-02 Source 1 · Source 2 · Source 3
ChatGPT / Codex / ChatGPT Pro 200USD 200/monthNot verifiedIndividual · standardMore usage than Pro 100 with no five-hour limit; weekly limits can apply; tasks remain metered. Includes $200/month of eligible API Platform credit after activation. Agent usage is shared. Task complexity and weekly limits affect capacity. GPT-5.5 retired from ChatGPT and Codex on October 14; Codex with ChatGPT stopped serving it at 19:00 UTC. API-key access is unaffected (official pricing and changelog, October 15). OpenAI canceled the Pro 200 reduction announced on September 29: new and existing Pro 200 subscriptions now get the higher Work and Codex allowance, and lower allowances doubled to match (help.openai.com, read 2026-11-01). Paid Pro accounts active for at least seven days can link an API organization they own using the same email to claim monthly API credits. The first grant is prorated; unused credit expires each billing cycle. API organization/model eligibility and rate limits apply. Credits cover eligible API Platform usage, excluding Codex with API keys, and do not increase signed-in Work or Codex limits.Documented · checked 2026-11-02 Source 1 · Source 2 · Source 3
Claude / Claude Code / Claude FreeFreeNot verifiedIndividual · standardLimited, variable usage. Chat and coding share usage. Absolute monthly capacity is undisclosed. With Opus 5.5 (anthropic.com, September 22, 2026) Anthropic raised five-hour limits on paid plans by an unpublished amount and granted a one-off limit reset; the announcement names no weekly-limit change.Documented · checked 2026-11-02 Source 1 · Source 2
Claude / Claude Code / Claude ProUSD 20/monthUSD 200Individual · standardVariable usage; shared chat and coding limits. Chat and coding share usage. Absolute monthly capacity is undisclosed. With Opus 5.5 (anthropic.com, September 22, 2026) Anthropic raised five-hour limits on paid plans by an unpublished amount and granted a one-off limit reset; the announcement names no weekly-limit change.Documented · checked 2026-11-02 Source 1 · Source 2
Claude / Claude Code / Claude MaxUSD 100/monthNot verifiedIndividual · standard5× Pro usage every five hours and every week, plus $100/month Claude API credit. Chat and coding share usage. Absolute monthly capacity is undisclosed. With Opus 5.5 (anthropic.com, September 22, 2026) Anthropic raised five-hour limits on paid plans by an unpublished amount and granted a one-off limit reset; the announcement names no weekly-limit change. Monthly API credits are rolling out, claimable after seven days on Max into one linked Console organization. They expire each billing cycle and cannot pay for interactive Claude Code or extra subscription usage.Documented · checked 2026-11-02 Source 1 · Source 2 · Source 3 · Source 4
Claude / Claude Code / Claude UltraUSD 200/monthNot verifiedIndividual · standard10× Pro usage every week with no five-hour limit for verified accounts, plus $200/month Claude API credit. Chat and coding share usage. Absolute monthly capacity is undisclosed. With Opus 5.5 (anthropic.com, September 22, 2026) Anthropic raised five-hour limits on paid plans by an unpublished amount and granted a one-off limit reset; the announcement names no weekly-limit change. Formerly Max 20x: existing subscribers moved to Ultra automatically at the same price (support.claude.com, read 2026-10-31). Until Anthropic verifies the account automatically, a temporary limit replaces the missing five-hour window. Monthly API credits are rolling out, claimable after seven days on Ultra into one linked Console organization. They expire each billing cycle and cannot pay for interactive Claude Code or extra subscription usage.Documented · checked 2026-11-02 Source 1 · Source 2 · Source 3 · Source 4
GitHub Copilot / Copilot FreeFreeNot verifiedIndividual · standard2,000 completions/month; limited AI credits. Credits are not tasks. Models consume them at different rates.Documented · checked 2026-11-01 Source 1 · Source 2
GitHub Copilot / Copilot ProUSD 10/monthNot verifiedIndividual · standard1,500 AI credits/month including flex allotment. Credits are not tasks. Models consume them at different rates.Documented · checked 2026-11-01 Source 1 · Source 2
GitHub Copilot / Copilot Pro+USD 39/monthNot verifiedIndividual · standard7,000 AI credits/month including flex allotment. Credits are not tasks. Models consume them at different rates. Fable 5.1 retains prompts and outputs for safety classifiers by default; an enterprise ZDR exemption requires separate approval and enablement.Documented · checked 2026-11-01 Source 1 · Source 2
GitHub Copilot / Copilot MaxUSD 100/monthNot verifiedIndividual · standard20,000 AI credits/month including flex allotment. Credits are not tasks. Models consume them at different rates. Fable 5.1 retains prompts and outputs for safety classifiers by default; an enterprise ZDR exemption requires separate approval and enablement. Copilot Fusion has a temporary 30% discount through November 30, 2026; the recurring model-specific value excludes it. Standard Auto selection has a 10% discount in the documented Chat/cloud surfaces.Documented · checked 2026-11-01 Source 1 · Source 2

For example, annual Claude Pro costs $200 upfront, equivalent to $16.67/month. No undisclosed allowance is treated as unlimited.

Explore model capability · Find models for your task

Interactive version: theaggregate.ai/subscriptions · How It Works · Data refreshed daily, snapshot 2026-11-05.