Chess Puzzles
Each model’s result as Epoch AI’s Benchmarking Hub publishes it, run by Epoch or collected from the source it cites. Facthory has not run them.
Data: Epoch AI, ‘Capabilities & benchmarking’, epoch.ai/benchmarks, licensed CC BY 4.0. Read on 16 Nov 2026.
The chart shows the 25 best and the table the 100 best of 164 results.
| # | Model | Organization | Result | Evidence |
|---|---|---|---|---|
| 1 | Claude Fable 5.2 (max) API access | Anthropic | 79% ± 4.1% | – |
| 2 | Gemini 4 Carbon (high) API access | Google DeepMind | 78% ± 4.2% | – |
| 3 | GPT-6 Astra (pro, max) API access | OpenAI | 76% ± 4.3% | Log or source |
| 4 | GPT-6 Astra (max) API access | OpenAI | 74% ± 4.4% | Log or source |
| 5 | GPT-5.6 Sol (pro, max) API access | OpenAI | 68% ± 4.7% | Log or source |
| 6 | GPT-6.1 Sol (max) API access | OpenAI | 68% ± 4.7% | Log or source |
| 7 | Claude Opus 5.5 (max) API access | Anthropic | 67% ± 4.7% | Log or source |
| 8 | Gemini 3.7 Flash (high) API access | Google DeepMind | 66% ± 4.8% | Log or source |
| 9 | GPT-5.5 (xhigh) API access | OpenAI | 63% ± 4.9% | Log or source |
| 10 | GPT-5.5 Pro (xhigh) API access | OpenAI | 62% ± 4.9% | Log or source |
| 11 | GPT-5.2 Pro API access | OpenAI | 60% ± 4.9% | Log or source |
| 12 | Claude Sonnet 5.5 (max) API access | Anthropic | 59% ± 4.9% | Log or source |
| 13 | Gemini 4 Argon (high) API access | Google DeepMind | 59% ± 4.9% | Log or source |
| 14 | GPT-5.4 Pro (xhigh) API access | OpenAI | 59% ± 4.9% | Log or source |
| 15 | GPT-5.6 Sol (max) API access | OpenAI | 59% ± 4.9% | Log or source |
| 16 | Gemini 3.8 Flash (high) API access | Google DeepMind | 58% ± 5% | Log or source |
| 17 | GPT-6 Sol (max) API access | OpenAI | 56% ± 5% | Log or source |
| 18 | Claude Haiku 5.5 (max) API access | Anthropic | 55% ± 5% | Log or source |
| 19 | Claude Opus 4.8 API access | Anthropic | 52% ± 5% | Log or source |
| 20 | Gemini 3.1 Pro Preview API access | Google DeepMind | 51% ± 5% | Log or source |
| 21 | Claude Fable 5.1 (max) API access | Anthropic | 48% ± 5% | Log or source |
| 22 | DeepSeek V4.1 Flash (max) Open weights (unrestricted) | DeepSeek | 48% ± 5% | Log or source |
| 23 | GPT-5 Pro API access | OpenAI | 48% ± 5% | Log or source |
| 24 | GPT-5.2 (xhigh) API access | OpenAI | 48% ± 5% | Log or source |
| 25 | GPT-5.4 (xhigh) API access | OpenAI | 48% ± 5% | Log or source |
| 26 | Claude Fable 5 (max) API access | Anthropic | 47% ± 5% | Log or source |
| 27 | GPT-5.6 Terra (max) API access | OpenAI | 47% ± 5% | Log or source |
| 28 | DeepSeek V4 Pro 0813 (max) Open weights (unrestricted) | DeepSeek | 46% ± 5% | Log or source |
| 29 | DeepSeek V4.1 Flash 1017 (max) Open weights (unrestricted) | DeepSeek | 46% ± 5% | Log or source |
| 30 | Qwen3.8 Max (xhigh) API access | Alibaba | 46% ± 5% | Log or source |
| 31 | Claude Fable 5 (high) API access | Anthropic | 45% ± 5% | Log or source |
| 32 | Claude Sonnet 5 (max) API access | Anthropic | 45% ± 5% | Log or source |
| 33 | Gemini 3.5 Flash API access | Google DeepMind | 45% ± 5% | Log or source |
| 34 | GPT-5.2 (high) API access | OpenAI | 45% ± 5% | Log or source |
| 35 | GPT-5.4 nano (xhigh) API access | OpenAI | 45% ± 5% | Log or source |
| 36 | Kimi K3.1 Pro (Together, max) Open weights (restricted use) | Moonshot | 45% ± 5% | Log or source |
| 37 | Claude Opus 5 (max) API access | Anthropic | 43% ± 5% | Log or source |
| 38 | GPT-5.6 Luna (max) API access | OpenAI | 43% ± 5% | Log or source |
| 39 | Gemini 3.5 Flash API access | Google DeepMind | 42% ± 5% | Log or source |
| 40 | Gemini 3.6 Flash API access | Google DeepMind | 42% ± 5% | Log or source |
| 41 | GPT-5.1 (high) API access | OpenAI | 42% ± 5% | Log or source |
| 42 | GPT-5.4 (high) API access | OpenAI | 42% ± 5% | Log or source |
| 43 | Claude Opus 4.7 (max) API access | Anthropic | 41% ± 4.9% | Log or source |
| 44 | Grok 4.7 (xhigh) API access | xAI | 41% ± 4.9% | Log or source |
| 45 | Doubao Seed Evolving (Nov 2026, high) API access | ByteDance | 40% ± 4.9% | Log or source |
| 46 | Gemini 3 Flash Preview API access | Google DeepMind | 40% ± 4.9% | Log or source |
| 47 | Gemini 3.6 Flash API access | Google DeepMind | 40% ± 4.9% | Log or source |
| 48 | GLM-5.3 (max) Open weights (restricted use) | Z.ai (Zhipu AI) | 37% ± 4.9% | Log or source |
| 49 | GPT-5.4 (medium) API access | OpenAI | 37% ± 4.9% | Log or source |
| 50 | Kimi K3 (high) Open weights (non-commercial) | Moonshot | 37% ± 4.9% | Log or source |
| 51 | Muse Spark 1.3 (max) API access | Meta AI | 37% ± 4.9% | Log or source |
| 52 | o3 (medium) API access | OpenAI | 37% ± 4.9% | Log or source |
| 53 | Qwen3.8 Max (0902) (xhigh) API access | Alibaba | 37% ± 4.9% | Log or source |
| 54 | Claude Opus 4.7 (xhigh) API access | Anthropic | 36% ± 4.8% | Log or source |
| 55 | GPT-5 (high) API access | OpenAI | 36% ± 4.8% | Log or source |
| 56 | GPT-5.2 (medium) API access | OpenAI | 36% ± 4.8% | Log or source |
| 57 | GPT-5.4 nano (high) API access | OpenAI | 36% ± 4.8% | Log or source |
| 58 | GPT-6 Luna (max) API access | OpenAI | 36% ± 4.8% | Log or source |
| 59 | Kimi K3 (max) Open weights (non-commercial) | Moonshot | 36% ± 4.8% | Log or source |
| 60 | o3 (high) API access | OpenAI | 36% ± 4.8% | Log or source |
| 61 | o3 (low) API access | OpenAI | 36% ± 4.8% | Log or source |
| 62 | Grok 4.5 (high) API access | xAI | 35% ± 4.8% | Log or source |
| 63 | Claude Sonnet 5 (xhigh) API access | Anthropic | 34% ± 4.8% | Log or source |
| 64 | Muse Spark 1.1 (xhigh) API access | Meta AI | 34% ± 4.8% | Log or source |
| 65 | Muse Spark 1.3 (xhigh) API access | Meta AI | 34% ± 4.8% | Log or source |
| 66 | Grok 4.6 (high) API access | xAI | 33% ± 4.7% | Log or source |
| 67 | Grok 4.6 (xhigh) API access | xAI | 33% ± 4.7% | Log or source |
| 68 | Mistral Large 4 Preview (high) Open weights (restricted use) | Mistral AI | 33% ± 4.7% | Log or source |
| 69 | DeepSeek V4 Flash 0731 (max) Open weights (unrestricted) | DeepSeek | 32% ± 4.7% | Log or source |
| 70 | GPT-5 (medium) API access | OpenAI | 32% ± 4.7% | Log or source |
| 71 | Claude Fable 5 (low) API access | Anthropic | 31% ± 4.6% | Log or source |
| 72 | GPT-5.6 Sol (low) API access | OpenAI | 31% ± 4.6% | Log or source |
| 73 | GPT-5.5 (low) API access | OpenAI | 30% ± 4.6% | Log or source |
| 74 | Claude Opus 4.8 API access | Anthropic | 29% ± 4.6% | Log or source |
| 75 | Gemini 3.6 Flash API access | Google DeepMind | 29% ± 4.6% | Log or source |
| 76 | grok-4.20-0309-reasoning API access | xAI | 29% ± 4.6% | Log or source |
| 77 | Mistral Large 4 (high) Open weights (restricted use) | Mistral AI | 29% ± 4.6% | Log or source |
| 78 | Claude Opus 5 (low) API access | Anthropic | 27% ± 4.5% | Log or source |
| 79 | Gemini 3.5 Flash API access | Google DeepMind | 27% ± 4.5% | Log or source |
| 80 | GPT-5 (low) API access | OpenAI | 27% ± 4.5% | Log or source |
| 81 | GPT-5.6 Terra (low) API access | OpenAI | 27% ± 4.5% | Log or source |
| 82 | Muse Spark 1.2 (xhigh) API access | Meta AI | 27% ± 4.5% | Log or source |
| 83 | Qwen3.7 Max API access | Alibaba | 27% ± 4.5% | Log or source |
| 84 | Claude Opus 4.7 (low) API access | Anthropic | 26% ± 4.4% | Log or source |
| 85 | Doubao Seed 2.1 Turbo (high) API access | ByteDance | 26% ± 4.4% | Log or source |
| 86 | GPT-5.2 (low) API access | OpenAI | 26% ± 4.4% | Log or source |
| 87 | Claude Haiku 5.5 (low) API access | Anthropic | 24% ± 4.3% | Log or source |
| 88 | GPT-5 mini (high) API access | OpenAI | 24% ± 4.3% | Log or source |
| 89 | Claude Sonnet 4.6 (max) API access | Anthropic | 23% ± 4.2% | Log or source |
| 90 | GPT-5.1 (low) API access | OpenAI | 23% ± 4.2% | Log or source |
| 91 | GPT-5.4 mini (high) API access | OpenAI | 23% ± 4.2% | Log or source |
| 92 | GLM-5.2 Open weights (unrestricted) | Z.ai (Zhipu AI) | 22% ± 4.2% | Log or source |
| 93 | GPT-5 nano (high) API access | OpenAI | 22% ± 4.2% | Log or source |
| 94 | o4-mini (medium) API access | OpenAI | 22% ± 4.2% | Log or source |
| 95 | Inkling (xhigh) Open weights (unrestricted) | Thinking Machines | 21% ± 4.1% | Log or source |
| 96 | Claude Opus 4.6 (max) API access | Anthropic | 21% ± 4.1% | Log or source |
| 97 | Doubao Seed 2.1 Pro (high) API access | ByteDance | 21% ± 4.1% | Log or source |
| 98 | Gemini 3.5 Flash-Lite API access | Google DeepMind | 21% ± 4.1% | Log or source |
| 99 | GLM-4.7 Open weights (unrestricted) | Z.ai (Zhipu AI) | 21% ± 4.1% | Log or source |
| 100 | GLM-5.3-Flash (max) Open weights (unrestricted) | Z.ai (Zhipu AI) | 21% ± 4.1% | Log or source |
