LLM Stats: leaderboard

Aggregated AI leaderboards comparing LLM, TTS, STT, video, image, and embedding models with performance, pricing, and capability data.

Source: llm-stats.com.

Benchmark results

Results are published separately for each benchmark or evaluation setting.

• LLM Stats (AA-Briefcase v1.1)
• LLM Stats (Agents' Last Exam)
• LLM Stats (AI2D)
• LLM Stats (AIME 2024)
• LLM Stats (AIME 2026)
• LLM Stats (AlignBench)
• LLM Stats (AMC_2022_23)
• LLM Stats (AndroidWorld_SR)
• LLM Stats (AndroidWorld)
• LLM Stats (APEX-Agents)
• LLM Stats (ARC-AGI-3)
• LLM Stats (Artificial Analysis)
• LLM Stats (ArXivMath)
• LLM Stats (AutomationBench v1.0.6)
• LLM Stats (AutomationBench-AA)
• LLM Stats (AutomationBench; API default fallbacks enabled)
• LLM Stats (AutomationBench; No fallback models; safeguards enabled)
• LLM Stats (AutomationBench; Private held-out set)
• LLM Stats (AutomationBench; Protocol unverified — claude-opus-5)
• LLM Stats (AutomationBench; Protocol unverified — deepseek-v4.1-flash)
• LLM Stats (AutomationBench; Protocol unverified — gpt-5.6-luna)
• LLM Stats (AutomationBench; Protocol unverified — gpt-5.6-sol)
• LLM Stats (AutomationBench; Protocol unverified — gpt-5.6-terra)
• LLM Stats (AutomationBench; Protocol unverified — gpt-6-astra)
• LLM Stats (AutomationBench; Protocol unverified — mistral-large-4)
• LLM Stats (AutomationBench; Protocol unverified — qwen3.8-max)
• LLM Stats (AutomationBench; Public (DeepSeek minimal harness))
• LLM Stats (AutomationBench; Public 600-task subset)
• LLM Stats (AutomationBench; Public v3)
• LLM Stats (AutomationBench; v1.0.6 method cohort)
• LLM Stats (BabyVision)
• LLM Stats (BenchCAD (with Python tool))
• LLM Stats (BenchCAD)
• LLM Stats (Beyond AIME)
• LLM Stats (BioMysteryBench)
• LLM Stats (Bird-SQL (dev))
• LLM Stats (BLINK)
• LLM Stats (BrowseComp Long Context 128k)
• LLM Stats (BrowseComp-zh)
• LLM Stats (BrowseComp)
• LLM Stats (C-Eval)
• LLM Stats (CC-OCR)
• LLM Stats (CharadesSTA)
• LLM Stats (Chartography)
• LLM Stats (ChartQA)
• LLM Stats (CharXiv-D)
• LLM Stats (CharXiv-R)
• LLM Stats (Claw-Eval)
• LLM Stats (ClawEval-MM)
• LLM Stats (CMMLU)
• LLM Stats (Codeforces - Qwen3.5 Unresolved Source Fraction)
• LLM Stats (Codeforces Rating - DeepSeek V3.1 Div1)
• LLM Stats (Codeforces Rating - DeepSeek V3.2 Exp report)
• LLM Stats (Codeforces Rating - DeepSeek V3.2 report)
• LLM Stats (Codeforces Rating - DeepSeek V4 family)
• LLM Stats (Codeforces Rating - DeepSeek V4.1 report)
• LLM Stats (Codeforces Rating - DiffusionGemma EntropyBound)
• LLM Stats (Codeforces Rating - Gemma4 unified)
• LLM Stats (Codeforces Rating - GPT-OSS no tools)
• LLM Stats (Codeforces Rating - Qwen3 CodeElo)
• LLM Stats (Codeforces Rating - Seed2 June-Dec2025)
• LLM Stats (COLLIE)
• LLM Stats (ComplexFuncBench)
• LLM Stats (CountBench)
• LLM Stats (CoWorkBench)
• LLM Stats (CritPT)
• LLM Stats (CSimpleQA)
• LLM Stats (CyberGym)
• LLM Stats (DeepPlanning)
• LLM Stats (DeepSearchQA)
• LLM Stats (DeepSWE 1.1)
• LLM Stats (DeepSWE)
• LLM Stats (DocVQA)
• LLM Stats (DocVQAtest)
• LLM Stats (DROP)
• LLM Stats (DynaMath)
• LLM Stats (ECLeKTic)
• LLM Stats (EgoSchema)
• LLM Stats (EmbSpatialBench)
• LLM Stats (ERQA)
• LLM Stats (ExploitBench)
• LLM Stats (ExploitGym)
• LLM Stats (Finance Agent v2)
• LLM Stats (Finance Agent)
• LLM Stats (FLEURS)
• LLM Stats (FrontierCode 1.1)
• LLM Stats (FrontierMath Tier 4 (v2))
• LLM Stats (FrontierScience Research)
• LLM Stats (FrontierSWE V2)
• LLM Stats (FrontierSWE)
• LLM Stats (GDP.pdf)
• LLM Stats (GDPval-AA 2.1)
• LLM Stats (GDPval-AA)
• LLM Stats (Graphwalks BFS <128k)
• LLM Stats (Graphwalks BFS >128k; 1M subset (F1))
• LLM Stats (Graphwalks BFS >128k; 256K subset (F1))
• LLM Stats (Graphwalks BFS >128k; 256K–1M range (F1))
• LLM Stats (Graphwalks BFS >128k; Protocol unverified — gpt-4.1-2025-04-14)
• LLM Stats (Graphwalks BFS >128k; Protocol unverified — gpt-4.1-mini-2025-04-14)
• LLM Stats (Graphwalks BFS >128k; Protocol unverified — gpt-4.1-nano-2025-04-14)
• LLM Stats (Graphwalks parents <128k)
• LLM Stats (Graphwalks parents >128k; 1M subset (F1))
• LLM Stats (Graphwalks parents >128k; 256K subset (F1))
• LLM Stats (Graphwalks parents >128k; 256K–1M range (Accuracy))
• LLM Stats (Graphwalks parents >128k; Protocol unverified — gpt-4.1-2025-04-14)
• LLM Stats (Graphwalks parents >128k; Protocol unverified — gpt-4.1-mini-2025-04-14)
• LLM Stats (Graphwalks parents >128k; Protocol unverified — gpt-4.1-nano-2025-04-14)
• LLM Stats (Hallusion Bench)
• LLM Stats (Harvey LAB-AA)
• LLM Stats (HealthBench Consensus)
• LLM Stats (HealthBench Hard)
• LLM Stats (HealthBench Professional)
• LLM Stats (HealthBench; Length-adjusted)
• LLM Stats (HealthBench; Protocol unverified — gpt-5.3-chat-latest)
• LLM Stats (HealthBench; Protocol unverified — kimi-k2-thinking-0905)
• LLM Stats (HealthBench; Protocol unverified — qwen3.8-max)
• LLM Stats (HealthBench; Protocol unverified — seed-2.0-mini)
• LLM Stats (HealthBench; Raw (GPT-OSS evaluation))
• LLM Stats (HiddenMath)
• LLM Stats (HMMT 2025)
• LLM Stats (HMMT Feb 26)
• LLM Stats (HMMT25; February 2025 (Granite evaluation))
• LLM Stats (HMMT25; November 2025)
• LLM Stats (HMMT25; Protocol unverified — grok-4-heavy)
• LLM Stats (HMMT25; Protocol unverified — grok-4)
• LLM Stats (HMMT25; Protocol unverified — qwen3-235b-a22b-instruct-2507)
• LLM Stats (HMMT25; Protocol unverified — qwen3-235b-a22b-thinking-2507)
• LLM Stats (HMMT25; Protocol unverified — qwen3-next-80b-a3b-instruct)
• LLM Stats (HMMT25; Protocol unverified — qwen3-next-80b-a3b-thinking)
• LLM Stats (HMMT25; Protocol unverified — qwen3-vl-235b-a22b-instruct)
• LLM Stats (HMMT25; Protocol unverified — qwen3-vl-235b-a22b-thinking)
• LLM Stats (HMMT25; Protocol unverified — qwen3-vl-30b-a3b-instruct)
• LLM Stats (HMMT25; Protocol unverified — qwen3-vl-30b-a3b-thinking)
• LLM Stats (HMMT25; Protocol unverified — qwen3-vl-4b-instruct)
• LLM Stats (HMMT25; Protocol unverified — qwen3-vl-4b-thinking)
• LLM Stats (HMMT25; Protocol unverified — qwen3-vl-8b-instruct)
• LLM Stats (HMMT25; Protocol unverified — qwen3-vl-8b-thinking)
• LLM Stats (Humanity's Last Exam (no tools, text-only))
• LLM Stats (Humanity's Last Exam (with tools, text-only))
• LLM Stats (IMO-AnswerBench)
• LLM Stats (Include)
• LLM Stats (InfoVQA)
• LLM Stats (InfoVQAtest)
• LLM Stats (Internal API instruction following (hard))
• LLM Stats (Job Bench)
• LLM Stats (JobBench)
• LLM Stats (Legal Agent Benchmark)
• LLM Stats (LifeSciBench)
• LLM Stats (LongBench v2)
• LLM Stats (LongVideoBench)
• LLM Stats (LVBench)
• LLM Stats (MathArena Apex)
• LLM Stats (MathVision)
• LLM Stats (MathVista-Mini)
• LLM Stats (MathVista)
• LLM Stats (MAXIFE)
• LLM Stats (MCP Atlas)
• LLM Stats (MCP-Mark)
• LLM Stats (MedXpertQA)
• LLM Stats (MLVU-M)
• LLM Stats (MLVU)
• LLM Stats (MM-MT-Bench)
• LLM Stats (MMAU)
• LLM Stats (MMBench-V1.1)
• LLM Stats (MMBench)
• LLM Stats (MMLongBench-Doc)
• LLM Stats (MMLU-ProX)
• LLM Stats (MMLU-Redux)
• LLM Stats (MMMLU)
• LLM Stats (MMStar)
• LLM Stats (MMVU)
• LLM Stats (MotionBench)
• LLM Stats (MRCR v2 (8-needle))
• LLM Stats (MRCR)
• LLM Stats (MuirBench)
• LLM Stats (Multi-Challenge)
• LLM Stats (Multi-IF)
• LLM Stats (Multi-SWE-Bench)
• LLM Stats (Multilingual MMLU)
• LLM Stats (MultiPL-E)
• LLM Stats (MVBench)
• LLM Stats (Natural Questions)
• LLM Stats (Natural2Code)
• LLM Stats (NL2Repo)
• LLM Stats (NOVA-63)
• LLM Stats (OCRBench_V2)
• LLM Stats (OCRBench-V2 (en))
• LLM Stats (OCRBench-V2 (zh))
• LLM Stats (OCRBench)
• LLM Stats (ODinW)
• LLM Stats (OfficeQA Pro)
• LLM Stats (OJBench)
• LLM Stats (OmniDocBench 1.5)
• LLM Stats (OmniScience)
• LLM Stats (OpenAI-MRCR: 2 needle 128k)
• LLM Stats (OpenAI-MRCR: 2 needle 1M)
• LLM Stats (OSWorld 2.0)
• LLM Stats (OSWorld 2.1 (partial))
• LLM Stats (OSWorld 2.1 (strict))
• LLM Stats (OSWorld-Verified)
• LLM Stats (OSWorld)
• LLM Stats (PinchBench - GLM 5V Turbo Launch Average)
• LLM Stats (PinchBench - Liquid LFM2.5 Reported)
• LLM Stats (PinchBench - NVIDIA Nemotron 3 Ultra Reported)
• LLM Stats (PinchBench - NVIDIA Nemotron 3.5 Bolt IQ2)
• LLM Stats (PinchBench - NVIDIA Nemotron 3.5 Lightning NeMo Gym)
• LLM Stats (PinchBench - Xiaomi MiMo V2 Launch Average)
• LLM Stats (PolyMATH)
• LLM Stats (PostTrainBench)
• LLM Stats (Program Bench)
• LLM Stats (RealWorldQA)
• LLM Stats (RefCOCO-avg)
• LLM Stats (RefSpatialBench)
• LLM Stats (RULER)
• LLM Stats (ScreenSpot Pro)
• LLM Stats (ScreenSpot)
• LLM Stats (Seal-0)
• LLM Stats (SEC-bench Pro)
• LLM Stats (SimpleQA Verified)
• LLM Stats (SimpleVQA)
• LLM Stats (SkillsBench)
• LLM Stats (Social IQa)
• LLM Stats (SQuALITY)
• LLM Stats (SuperChem)
• LLM Stats (SWE Atlas - Codebase QnA)
• LLM Stats (SWE-bench Multilingual)
• LLM Stats (SWE-Bench Multimodal)
• LLM Stats (SWE-Lancer (IC-Diamond subset))
• LLM Stats (SWE-Marathon)
• LLM Stats (t2-bench)
• LLM Stats (TAU-bench Airline; Claude extended thinking; prompt addendum; 100 steps)
• LLM Stats (TAU-bench Airline; GLM optimized user simulator)
• LLM Stats (TAU-bench Airline; GPT-4o user; five runs; no custom prompting)
• LLM Stats (TAU-bench Airline; Protocol unverified — claude-3-5-haiku-20241022)
• LLM Stats (TAU-bench Airline; Protocol unverified — claude-3-5-sonnet-20241022)
• LLM Stats (TAU-bench Airline; Protocol unverified — claude-3-7-sonnet-20250219)
• LLM Stats (TAU-bench Airline; Protocol unverified — claude-sonnet-4-5-20250929)
• LLM Stats (TAU-bench Airline; Protocol unverified — gpt-4.5)
• LLM Stats (TAU-bench Airline; Protocol unverified — gpt-4o-2024-08-06)
• LLM Stats (TAU-bench Airline; Protocol unverified — minimax-m1-40k)
• LLM Stats (TAU-bench Airline; Protocol unverified — minimax-m1-80k)
• LLM Stats (TAU-bench Airline; Protocol unverified — o1-2024-12-17)
• LLM Stats (TAU-bench Airline; Protocol unverified — o3-mini)
• LLM Stats (TAU-bench Airline; Protocol unverified — o4-mini)
• LLM Stats (TAU-bench Airline; Protocol unverified — qwen3-235b-a22b-thinking-2507)
• LLM Stats (TAU-bench Airline; Protocol unverified — qwen3-coder-480b-a35b-instruct)
• LLM Stats (TAU-bench Airline; Protocol unverified — qwen3-next-80b-a3b-instruct)
• LLM Stats (TAU-bench Airline; Protocol unverified — qwen3-next-80b-a3b-thinking)
• LLM Stats (TAU-bench Retail; Claude extended thinking; prompt addendum; 100 steps)
• LLM Stats (TAU-bench Retail; GLM optimized user simulator)
• LLM Stats (TAU-bench Retail; GPT-4o user; five runs; no custom prompting)
• LLM Stats (TAU-bench Retail; Protocol unverified — claude-3-5-haiku-20241022)
• LLM Stats (TAU-bench Retail; Protocol unverified — claude-3-5-sonnet-20241022)
• LLM Stats (TAU-bench Retail; Protocol unverified — claude-3-7-sonnet-20250219)
• LLM Stats (TAU-bench Retail; Protocol unverified — claude-sonnet-4-5-20250929)
• LLM Stats (TAU-bench Retail; Protocol unverified — gpt-4.5)
• LLM Stats (TAU-bench Retail; Protocol unverified — gpt-4o-2024-08-06)
• LLM Stats (TAU-bench Retail; Protocol unverified — gpt-oss-120b)
• LLM Stats (TAU-bench Retail; Protocol unverified — gpt-oss-20b)
• LLM Stats (TAU-bench Retail; Protocol unverified — minimax-m1-40k)
• LLM Stats (TAU-bench Retail; Protocol unverified — minimax-m1-80k)
• LLM Stats (TAU-bench Retail; Protocol unverified — o1-2024-12-17)
• LLM Stats (TAU-bench Retail; Protocol unverified — o3-mini)
• LLM Stats (TAU-bench Retail; Protocol unverified — o4-mini)
• LLM Stats (TAU-bench Retail; Protocol unverified — qwen3-235b-a22b-thinking-2507)
• LLM Stats (TAU-bench Retail; Protocol unverified — qwen3-coder-480b-a35b-instruct)
• LLM Stats (TAU-bench Retail; Protocol unverified — qwen3-next-80b-a3b-instruct)
• LLM Stats (TAU-bench Retail; Protocol unverified — qwen3-next-80b-a3b-thinking)
• LLM Stats (Tau3 Banking)
• LLM Stats (TAU3-Bench)
• LLM Stats (Terminal-Bench 2.1)
• LLM Stats (Terminal-Bench 4.0)
• LLM Stats (Terminal-Bench-Science 0.1)
• LLM Stats (TextVQA)
• LLM Stats (TheoremQA)
• LLM Stats (Toolathlon - Verified protocol)
• LLM Stats (Toolathlon-Verified)
• LLM Stats (Toolathlon)
• LLM Stats (TriviaQA)
• LLM Stats (TVBench)
• LLM Stats (V*)
• LLM Stats (Vals Finance Agent v2)
• LLM Stats (Vibe-Eval)
• LLM Stats (Video-MME)
• LLM Stats (VideoMME w sub.)
• LLM Stats (VideoMME w/o sub.)
• LLM Stats (VideoMMMU)
• LLM Stats (Vision2Web)
• LLM Stats (VITA-Bench)
• LLM Stats (VLMsAreBlind)
• LLM Stats (WideSearch)
• LLM Stats (Wild Bench)
• LLM Stats (WildClawBench)
• LLM Stats (WMT24++)
• LLM Stats (WorldVQA)
• LLM Stats (WritingBench)
• LLM Stats (ZebraLogic)
• LLM Stats (ZEROBench-Sub)
• LLM Stats (ZEROBench)
• LLM Stats Score

Interactive version: theaggregate.ai/benchmark?slug=llm-stats · How It Works · Data refreshed daily, snapshot 2026-10-20.