Epoch AI - Weirdml V3: leaderboard
WeirdML v3 as run by Epoch AI: agentic tasks in which a coding agent explores unfamiliar data and builds machine-learning pipelines with unspecified goals and limited feedback. The score combines the area under the score-versus-tokens curve with the final value; higher is better. Not comparable with v2.
Metric: Score. Source: epoch.ai. Saturation forecast: Around December 2027. 26 models tracked. Updated 10 Nov 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | GPT-6 Astra | xHigh | 42.22 | #6 |
| 2 | GPT-6.1 Sol | xHigh | 38.25 | #10 |
| 3 | Claude Opus 5.5 | xHigh | 31.2 | #5 |
| 4 | GPT-6.1 Terra | xHigh | 28.09 | #21 |
| 5 | Claude Fable 5.1 | xHigh | 25.97 | #15 |
| 6 | Gemini 4 Argon | High | 20.26 | #12 |
| 7 | Claude Sonnet 5.5 | xHigh | 19.94 | #19 |
| 8 | GPT-6 Sol | xHigh | 19.73 | #26 |
| 9 | Claude Opus 5 | xHigh | 19.01 | #20 |
| 10 | Claude Fable 5 | xHigh | 15.59 | #27 |
| 11 | muse-wonder-1.0 | xHigh | 15.17 | |
| 12 | GPT-5.6 Sol | xHigh | 14.99 | #30 |
| 13 | Claude Haiku 5.5 | xHigh | 11.7 | #70 |
| 14 | Kimi K3.1 Pro | xHigh | 9.89 | #29 |
| 15 | DeepSeek V4.1 Flash (1017) | High | 9.19 | #41 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon.
Interactive version: theaggregate.ai/benchmark?slug=epoch-ai-weirdml-v3 · How It Works · Data refreshed daily, snapshot 2026-11-16.