Chess Puzzles (task version 2.0)
Games · 2026-10-13
Epoch AI's Chess Puzzles reruns under task version 2.0.x, published October 13-14, 2026. Epoch replaced most models' 1.x runs with these and published no changelog; several scores moved sharply (e.g. Claude Opus 4.7 max from 7 to 41), so the protocol is treated as distinct from 1.x. It is its own capability family, separate from the 1.x family: 80 models have 2.0 results and only 7 also have 1.x results, so the two protocols enter the fit side by side rather than pooled.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-6 Astra Pro | 76.0 |
| GPT-6 Astra | 74.0 |
| GPT-5.6 Sol Pro | 68.0 |
| GPT-6.1 Sol | 68.0 |
| Claude Opus 5.5 | 67.0 |
| Gemini 3.7 Flash | 66.0 |
| GPT-5.5 | 63.0 |
| GPT-5.5 Pro | 62.0 |
| GPT-5.2 Pro | 60.0 |
| Claude Sonnet 5.5 | 59.0 |
| Gemini 4 Argon | 59.0 |
| GPT-5.4 Pro | 59.0 |
| GPT-5.6 Sol | 59.0 |
| Gemini 3.8 Flash | 58.0 |
| GPT-6 Sol | 56.0 |
| Claude Haiku 5.5 | 55.0 |
| Opus 4.8 | 52.0 |
| Gemini 3.1 Pro Preview | 51.0 |
| Claude Fable 5.1 | 48.0 |
| DeepSeek V4.1 Flash | 48.0 |
| GPT-5.2 | 48.0 |
| GPT-5.4 | 48.0 |
| GPT-5 Pro | 48.0 |
| Claude Fable 5 | 47.0 |
| GPT-5.6 Terra | 47.0 |
| DeepSeek V4.1 Flash 1017 | 46.0 |
| DeepSeek V4 Pro 0813 | 46.0 |
| Sonnet 5 | 45.0 |
| Gemini 3.5 Flash | 45.0 |
| GPT-5.4 Nano | 45.0 |
| Claude Opus 5 | 43.0 |
| GPT-5.6 Luna | 43.0 |
| Gemini 3.6 Flash | 42.0 |
| GPT-5.1 | 42.0 |
| Opus 4.7 | 41.0 |
| Grok 4.7 | 41.0 |
| Gemini 3 Flash Preview | 39.0 |
| GLM-5.3 | 37.0 |
| Kimi K3 | 37.0 |
| Muse Spark 1.3 | 37.0 |