Benchmarks
SimpleBench
Models scored
91
Best score
0.82
Metric
Score (AVG@5)
SimpleBench is an AI evaluation tracked by Epoch AI, with 91 scored model versions on record. The scores here are reported by third parties rather than produced by Epoch AI running the evaluation itself.
Scored models span February 2024 to July 2026. The best score recorded is 0.82, measured as Score (AVG@5). Models come from Anthropic, Google DeepMind, OpenAI, Google, Alibaba, xAI, DeepSeek, Moonshot and others.
Full record
- Score metric
- Score (AVG@5)
- Models scored
- 91
- Best score recorded
- 0.82
- Earliest model scored
- Feb 29, 2024
- Most recent model scored
- Jul 9, 2026
- Third-party reported
- Yes
Leaderboard
| 01 | claude-fable-5_max | Anthropic | 0.82 |
| 02 | gemini-3.1-pro-preview | Google DeepMind | 0.8 |
| 03 | gpt-5.5-pro_high | OpenAI | 0.77 |
| 04 | gpt-5.5-pro_unknown | OpenAI | 0.77 |
| 05 | gemini-3.5-flash_unknown | 0.77 | |
| 06 | gemini-3-pro-preview | Google DeepMind | 0.76 |
| 07 | gpt-5.4-pro-2026-03-05_unknown | OpenAI | 0.74 |
| 08 | gpt-5.6-sol_prounknown | OpenAI | 0.72 |
| 09 | qwen3.7-max | Alibaba | 0.7 |
| 10 | grok-4.5_unknown | xAI | 0.7 |
| 11 | gpt-5.5_unknown | OpenAI | 0.69 |
| 12 | claude-opus-4-6 | Anthropic | 0.68 |
| 13 | claude-opus-4-8_unknown | Anthropic | 0.65 |
| 14 | gpt-5.6-sol_unknown | OpenAI | 0.65 |
| 15 | qwen3.6-max-preview | Alibaba | 0.63 |
| 16 | claude-opus-4-7_unknown | Anthropic | 0.63 |
| 17 | gemini-2.5-pro-preview-06-05 | Google DeepMind | 0.62 |
| 18 | claude-opus-4-5-20251101_unknown | Anthropic | 0.62 |
| 19 | claude-opus-4-5-20251101 | Anthropic | 0.62 |
| 20 | gpt-5-pro-2025-10-06_unknown | OpenAI | 0.62 |
| 21 | gpt-5-pro-2025-10-06_high | OpenAI | 0.62 |
| 22 | deepseek-v4-pro_unknown | DeepSeek | 0.61 |
| 23 | gemini-3-flash-preview | Google DeepMind | 0.61 |
| 24 | claude-sonnet-5_unknown | Anthropic | 0.61 |
| 25 | grok-4-0709 | xAI | 0.6 |
| 26 | claude-opus-4-1-20250805 | Anthropic | 0.6 |
| 27 | claude-opus-4-1-20250805_unknown | Anthropic | 0.6 |
| 28 | claude-opus-4-20250514_unknown | Anthropic | 0.59 |
| 29 | claude-opus-4-20250514_12K | Anthropic | 0.59 |
| 30 | kimi-k2.7-code | Moonshot | 0.58 |
| 31 | gpt-5.2-pro-2025-12-11 | OpenAI | 0.57 |
| 32 | gpt-5-2025-08-07_high | OpenAI | 0.57 |
| 33 | grok-4-1-fast-non-reasoning | xAI | 0.56 |
| 34 | grok-4-1-fast-reasoning | xAI | 0.56 |
| 35 | glm-5.1 | Z.ai (Zhipu AI) | 0.55 |
| 36 | claude-sonnet-4-5-20250929_12K | Anthropic | 0.54 |
| 37 | claude-sonnet-4-5-20250929_unknown | Anthropic | 0.54 |
| 38 | glm-5 | Z.ai (Zhipu AI) | 0.53 |
| 39 | gpt-5.1-2025-11-13_high | OpenAI | 0.53 |
| 40 | o3-2025-04-16_high | OpenAI | 0.53 |
| 41 | DeepSeek-V3.2-Speciale | DeepSeek | 0.53 |
| 42 | gemini-2.5-pro-exp-03-25 | Google DeepMind | 0.52 |
| 43 | gemini-2.5-pro-preview-03-25 | Google DeepMind | 0.52 |
| 44 | gpt-5.6-terra_unknown | OpenAI | 0.49 |
| 45 | zai-org/glm-4.7 | Z.ai (Zhipu AI) | 0.48 |
| 46 | glm-4.7 | Z.ai (Zhipu AI) | 0.48 |
| 47 | gpt-5.6-luna_unknown | OpenAI | 0.47 |
| 48 | kimi-k2.5 | Moonshot | 0.47 |
| 49 | claude-3-7-sonnet-20250219_12K | Anthropic | 0.46 |
| 50 | claude-3-7-sonnet-20250219_unknown | Anthropic | 0.46 |
| 51 | gpt-5.2-2025-12-11_high | OpenAI | 0.46 |
| 52 | MiniMax-M3 | MiniMax | 0.46 |
| 53 | gpt-5.2-2025-12-11_unknown | OpenAI | 0.46 |
| 54 | claude-sonnet-4-20250514_unknown | Anthropic | 0.46 |
| 55 | claude-sonnet-4-20250514_12K | Anthropic | 0.46 |
| 56 | claude-3-7-sonnet-20250219 | Anthropic | 0.45 |
| 57 | o1-preview-2024-09-12 | OpenAI | 0.42 |
| 58 | claude-3-5-sonnet-20241022 | Anthropic | 0.41 |
| 59 | gemini-2.5-flash | Google DeepMind | 0.41 |
| 60 | DeepSeek-R1-0528 | DeepSeek | 0.41 |
More benchmarks
SourceEpoch AI, 'AI Benchmarking Hub'. Published online at epoch.ai. Retrieved 2026-07-29 from https://epoch.ai/benchmarks. Licensed under CC BY 4.0.
← All benchmarks