AA MKA - Climb WSR: leaderboard
Artificial Analysis MKA, 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 2029. 21 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 4 Carbon (High) | 54.34 | #2 (Gemini 4 Carbon) |
| 2 | Gemini 4 Argon (High) | 51.26 | #11 (Gemini 4 Argon) |
| 3 | GPT-6 Astra (Max) | 50.24 | #6 (GPT-6 Astra) |
| 4 | GPT-6.1 Sol (Max) | 47.72 | #10 (GPT-6.1 Sol) |
| 5 | Muse Wonder (Max) | 47.49 | #27 (Muse Wonder) |
| 6 | Kimi K3.1 Pro | 43.86 | #28 |
| 7 | GPT-6.1 Terra (Max) | 43.76 | #19 (GPT-6.1 Terra) |
| 8 | Claude Haiku 5.5 (Max) | 40.96 | #64 (Claude Haiku 5.5) |
| 9 | Grok 4.7 (xHigh) | 40.14 | #43 (Grok 4.7) |
| 10 | DeepSeek V4.1 Flash (1017) (Max) | 38.54 | #38 (DeepSeek V4.1 Flash (1017)) |
| 11 | Qwen 3.8 Max (0902) | 37.43 | #49 |
| 12 | MiMo-V2.6-Pro | 37.32 | #68 |
| 13 | GPT-6 Luna (Max) | 36.75 | #93 (GPT-6 Luna) |
| 14 | GLM-5.3 Flash | 36.24 | #119 |
| 15 | Gemini 3.8 Flash (High) | 35.27 | #36 (Gemini 3.8 Flash) |
No result here: #1 Claude Fable 5.2, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=aa-mka-climb-wsr · How It Works · Data refreshed daily, snapshot 2026-11-12.