games benchmark · included in ECI
Chess Puzzles
Chess tactics puzzles solved by the model without tools.
Results
65
Random baseline
5.0%
Score ceiling
100%
Released
10 Dec 2025
Score by model release date
Best score by organisation
Leaderboard
| # | Model | Score | Relative | Setting | Released | Run | Evidence |
|---|---|---|---|---|---|---|---|
| 1 | DeepSeek V4 Pro 0813 Open source DeepSeek | 47.0% ±5.0 | max | 13 Aug 2026 | 18 Aug 2026 | Source ↗ | |
| 2 | Kimi K3 Open source MoonshotAI | 39.0% ±4.9 | max | 16 Jul 2026 | 16 Jul 2026 | Source ↗ | |
| 3 | DeepSeek V4 Flash 0731 Open source DeepSeek | 33.0% ±4.7 | max | 31 Jul 2026 | 2 Aug 2026 | Source ↗ | |
| 4 | Qwen3.6 35B A3B Open source Qwen | 26.0% ±4.4 | — | 27 Apr 2026 | 7 Aug 2026 | Log ↗ | |
| 5 | Kimi K2.6 Open source MoonshotAI | 26.0% ±4.4 | — | 20 Apr 2026 | 7 May 2026 | Log ↗ | |
| 6 | Qwen3.6 27B Open source Qwen | 22.0% ±4.2 | — | 22 Apr 2026 | 7 Aug 2026 | Log ↗ | |
| 7 | GLM 5.3 Open source Z.ai | 21.0% ±4.1 | max | 14 Aug 2026 | 24 Aug 2026 | Source ↗ | |
| 8 | Inkling (xhigh) Open source Thinking Machines | 21.0% ±4.1 | xhigh | 15 Jul 2026 | 5 Aug 2026 | Source ↗ | |
| 9 | GLM 5.2 Open source Z.ai | 21.0% ±4.1 | max | 16 Jun 2026 | 17 Jun 2026 | Source ↗ | |
| 10 | Kimi K2.7 Code Open source MoonshotAI | 21.0% ±4.1 | — | 12 Jun 2026 | 7 Aug 2026 | Log ↗ | |
| 11 | Kimi K3 (low) Open source Moonshot | 20.0% ±4.0 | low | 16 Jul 2026 | 7 Aug 2026 | Log ↗ | |
| 12 | DeepSeek v4 Pro (high) Open source DeepSeek | 20.0% ±4.0 | max | 24 Apr 2026 | 16 Jun 2026 | Source ↗ | |
| 13 | Kimi K2 Thinking Turbo Open source Moonshot | 20.0% ±4.0 | — | 6 Nov 2025 | 10 Dec 2025 | Log ↗ | |
| 14 | gpt-oss-120b Open source OpenAI | 20.0% ±3.1 | high | 5 Aug 2025 | 11 Dec 2025 | Log ↗ | |
| 15 | GLM 5.1 Open source Z.ai | 19.0% ±3.9 | — | 7 Apr 2026 | 10 Aug 2026 | Log ↗ | |
| 16 | Inkling Small (xhigh) Open source Thinking Machines | 18.0% ±3.9 | xhigh | 15 Jul 2026 | 14 Aug 2026 | Source ↗ | |
| 17 | GLM 5.3 Flash Open source Z.ai | 14.0% ±3.5 | max | 20 Aug 2026 | 26 Aug 2026 | Source ↗ | |
| 18 | MiniMax M3 Open source MiniMax | 14.0% ±3.5 | — | 1 Jun 2026 | 10 Aug 2026 | Log ↗ | |
| 19 | DeepSeek V3.2 Open source DeepSeek | 14.0% ±3.5 | — | 1 Dec 2025 | 16 Dec 2025 | Log ↗ | |
| 20 | Qwen3.5 397B A17B Open source Qwen | 13.0% ±3.4 | none | 13 Feb 2026 | 7 Aug 2026 | Log ↗ | |
| 21 | Nemotron 3 Ultra Open source NVIDIA | 12.0% ±3.3 | — | 4 Jun 2026 | 10 Aug 2026 | Log ↗ | |
| 22 | Qwen3.5-9B Open source Qwen | 12.0% ±3.3 | none | 24 Feb 2026 | 27 Aug 2026 | Log ↗ | |
| 23 | Kimi K2.5 (Fireworks) Open source Moonshot | 12.0% ±3.3 | — | 27 Jan 2026 | 28 Jan 2026 | Log ↗ | |
| 24 | Qwen3-235B-A22B-Thinking (Jul 2025) Open source Alibaba | 12.0% ±3.3 | — | 25 Jul 2025 | 11 Dec 2025 | Log ↗ | |
| 25 | Qwen3.5-35B-A3B Open source Qwen | 10.0% ±3.0 | none | 24 Feb 2026 | 30 Aug 2026 | Log ↗ | |
| 26 | GLM 5 Open source Z.ai | 10.0% ±3.0 | — | 11 Feb 2026 | 12 Feb 2026 | Log ↗ | |
| 27 | Qwen3-30B-A3B-Thinking (Jul 2025) Open source Alibaba | 8.0% ±2.7 | — | 30 Jul 2025 | 30 Aug 2026 | Log ↗ | |
| 28 | DeepSeek v4 Pro (unknown thinking) Open source DeepSeek | 7.0% ±2.6 | none | 24 Apr 2026 | 6 Aug 2026 | Log ↗ | |
| 29 | Gemma 4 26B A4B Open source Google | 6.0% ±2.4 | minimal | 2 Apr 2026 | 6 Aug 2026 | Log ↗ | |
| 30 | GLM 4.7 Open source Z.ai | 6.0% ±1.6 | — | 22 Dec 2025 | 29 Jan 2026 | Log ↗ | |
| 31 | Gemma 4 31B Open source Google | 5.0% ±2.2 | minimal | 2 Apr 2026 | 6 Aug 2026 | Log ↗ | |
| 32 | Qwen3 32B Open source Qwen | 5.0% ±2.2 | — | 29 Apr 2025 | 28 Aug 2026 | Log ↗ | |
| 33 | Qwen3 8B Open source Qwen | 5.0% ±2.2 | — | 28 Apr 2025 | 27 Aug 2026 | Log ↗ | |
| 34 | QwQ-32B Open source Alibaba | 5.0% ±2.2 | — | 5 Mar 2025 | 28 Aug 2026 | Log ↗ | |
| 35 | gpt-oss-20b Open source OpenAI | 4.0% ±2.0 | medium | 5 Aug 2025 | 27 Aug 2026 | Log ↗ | |
| 36 | Qwen3 14B Open source Qwen | 4.0% ±2.0 | — | 29 Apr 2025 | 30 Aug 2026 | Log ↗ | |
| 37 | Qwen3 30B A3B Open source Qwen | 4.0% ±2.0 | — | 29 Apr 2025 | 30 Aug 2026 | Log ↗ | |
| 38 | Magistral Small 1.2 Open source Mistral AI | 3.0% ±1.7 | — | 18 Sep 2025 | 30 Aug 2026 | Log ↗ | |
| 39 | Qwen3-30B-A3B-Instruct (Jul 2025) Open source Alibaba | 2.0% ±1.4 | — | 29 Jul 2025 | 30 Aug 2026 | Log ↗ | |
| 40 | Mistral Small 3.2 Open source Mistral AI | 1.0% ±1.0 | — | 20 Jun 2025 | 30 Aug 2026 | Log ↗ | |
| 41 | Qwen3-4B Open source Alibaba | 1.0% ±1.0 | none | 29 Apr 2025 | 28 Aug 2026 | Log ↗ | |
| 42 | Mistral Small 3.1 Open source Mistral AI | 1.0% ±1.0 | — | 17 Mar 2025 | 30 Aug 2026 | Log ↗ | |
| 43 | DeepSeek-R1-Distill-Qwen-32B Open source DeepSeek | 1.0% ±1.0 | — | 20 Jan 2025 | 28 Aug 2026 | Log ↗ | |
| 44 | DeepSeek-R1-Distill-Qwen-14B Open source DeepSeek | 1.0% ±1.0 | — | 20 Jan 2025 | 28 Aug 2026 | Log ↗ | |
| 45 | Phi 4 Open source Microsoft | 1.0% ±1.0 | — | 12 Dec 2024 | 28 Aug 2026 | Log ↗ | |
| 46 | Qwen3.5-2B Open source Alibaba | 0.0% | none | 24 Feb 2026 | 28 Aug 2026 | Log ↗ | |
| 47 | GLM 4.7 Flash Open source Z.ai | 0.0% | none | 19 Jan 2026 | 28 Aug 2026 | Log ↗ | |
| 48 | Granite 4.0 Micro Open source IBM | 0.0% | — | 20 Oct 2025 | 30 Aug 2026 | Log ↗ | |
| 49 | Qwen3-1.7B Open source Alibaba | 0.0% | none | 29 Apr 2025 | 28 Aug 2026 | Log ↗ | |
| 50 | Gemma 3 4B Open source Google | 0.0% | — | 12 Mar 2025 | 28 Aug 2026 | Log ↗ | |
| 51 | Gemma 3 12B Open source Google | 0.0% | — | 12 Mar 2025 | 28 Aug 2026 | Log ↗ | |
| 52 | Gemma 3 27B Open source Google | 0.0% | — | 12 Mar 2025 | 27 Aug 2026 | Log ↗ | |
| 53 | Gemma 3 1B Open source Google DeepMind | 0.0% | — | 12 Mar 2025 | 28 Aug 2026 | Log ↗ | |
| 54 | Mistral Small 3 Open source Mistral | 0.0% | — | 30 Jan 2025 | 30 Aug 2026 | Log ↗ | |
| 55 | DeepSeek-R1-Distill-Qwen-1.5B Open source DeepSeek | 0.0% | — | 20 Jan 2025 | 28 Aug 2026 | Log ↗ | |
| 56 | Llama 3.2 1B Open source Meta AI | 0.0% | — | 24 Sep 2024 | 30 Aug 2026 | Log ↗ | |
| 57 | Qwen2.5-7B Open source Alibaba | 0.0% | — | 19 Sep 2024 | 30 Aug 2026 | Log ↗ | |
| 58 | Qwen2.5-32B Open source Alibaba | 0.0% | — | 17 Sep 2024 | 30 Aug 2026 | Log ↗ | |
| 59 | Llama 3.1-8B Open source Meta AI | 0.0% | — | 23 Jul 2024 | 27 Aug 2026 | Log ↗ | |
| 60 | Mistral 7B v0.3 Open source Mistral AI | 0.0% | — | 27 May 2024 | 30 Aug 2026 | Log ↗ | |
| 61 | phi-3-mini 3.8B Open source Microsoft | 0.0% | — | 23 Apr 2024 | 30 Aug 2026 | Log ↗ | |
| 62 | Llama 3-8B Open source Meta AI | 0.0% | — | 18 Apr 2024 | 28 Aug 2026 | Log ↗ | |
| 63 | DeepSeek LLM 67B Open source DeepSeek | 0.0% | — | 29 Nov 2023 | 28 Aug 2026 | Log ↗ | |
| 64 | Llama 2-13B Open source Meta AI | 0.0% | — | 18 Jul 2023 | 28 Aug 2026 | Log ↗ | |
| 65 | Llama 2-7B Open source Meta AI | 0.0% | — | 18 Jul 2023 | 30 Aug 2026 | Log ↗ |
Caveats: settings (reasoning effort, agent scaffold) differ between rows and materially affect scores; the best reported setting per model is shown. Source: Epoch AI, CC BY 4.0.