AA MKA - Vertical - Finance - Climb WSR: leaderboard
Artificial Analysis MKA, Vertical - Finance - Climb WSR: criterion-weighted task satisfaction, reported as a percentage. 200 private MKA tasks; one run per task, Stirrup, 250 turns and 8 hours; Linux sandbox with code, internet, search, page fetch and image viewing. GPT-6 Luna medium agentic judge; criterion weights 1–5, task-mean weighted satisfaction. Climb and Regression criterion populations remain separate (Regression covers 180 tasks). Empty submissions and errored runs score zero; final-text-only submissions are graded. The judge has no network or input files. Native effort and fallback labels retained. Source: https://artificialanalysis.ai/evaluations/mka
Metric: Weighted Satisfaction Rate (%). Source: artificialanalysis.ai. Saturation forecast: Around October 2028. 21 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 4 Argon (High) | 56.5 | #11 (Gemini 4 Argon) |
| 2 | Gemini 4 Carbon (High) | 55.59 | #2 (Gemini 4 Carbon) |
| 3 | Muse Wonder (Max) | 53.56 | #27 (Muse Wonder) |
| 4 | GPT-6 Astra (Max) | 51.18 | #6 (GPT-6 Astra) |
| 5 | Kimi K3.1 Pro | 50.25 | #28 |
| 6 | GPT-6.1 Terra (Max) | 47.29 | #19 (GPT-6.1 Terra) |
| 7 | GPT-6.1 Sol (Max) | 47.24 | #10 (GPT-6.1 Sol) |
| 8 | Claude Haiku 5.5 (Max) | 43.76 | #64 (Claude Haiku 5.5) |
| 9 | Grok 4.7 (xHigh) | 42.93 | #43 (Grok 4.7) |
| 10 | Qwen 3.8 Max (0902) | 41.54 | #49 |
| 11 | DeepSeek V4.1 Flash (1017) (Max) | 40.09 | #38 (DeepSeek V4.1 Flash (1017)) |
| 12 | MiMo-V2.6-Pro | 37 | #68 |
| 13 | GPT-6 Luna (Max) | 35.97 | #93 (GPT-6 Luna) |
| 14 | Gemini 3.8 Flash (High) | 35.03 | #36 (Gemini 3.8 Flash) |
| 15 | Mistral Large 4 (High) | 34.08 | #125 (Mistral Large 4) |
No result here: #1 Claude Fable 5.2, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=aa-mka-vertical-finance-climb-wsr · How It Works · Data refreshed daily, snapshot 2026-11-12.