Benchmarks
Critpt
Models scored
74
Best score
0.32
Metric
Accuracy
Critpt is an AI evaluation tracked by Epoch AI, with 74 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 July 2024 to July 2026. The best score recorded is 0.32, measured as Accuracy. Models come from OpenAI, Anthropic, Moonshot, Z.ai (Zhipu AI), Google DeepMind, Meta AI, Alibaba, Google and others.
Full record
- Score metric
- Accuracy
- Models scored
- 74
- Best score recorded
- 0.32
- Earliest model scored
- Jul 23, 2024
- Most recent model scored
- Jul 24, 2026
- Third-party reported
- Yes
Leaderboard
| 01 | gpt-5.6-sol_max | OpenAI | 0.32 |
| 02 | gpt-5.5-pro_xhigh | OpenAI | 0.31 |
| 03 | gpt-5.6-terra_max | OpenAI | 0.3 |
| 04 | gpt-5.4-pro-2026-03-05_xhigh | OpenAI | 0.3 |
| 05 | claude-opus-5_max | Anthropic | 0.29 |
| 06 | gpt-5.6-sol_xhigh | OpenAI | 0.29 |
| 07 | claude-fable-5_max | Anthropic | 0.29 |
| 08 | gpt-5.5_xhigh | OpenAI | 0.27 |
| 09 | gemini-3-deep-think-preview | 0.26 | |
| 10 | gpt-5.5_high | OpenAI | 0.25 |
| 11 | gpt-5.4-2026-03-05_xhigh | OpenAI | 0.23 |
| 12 | kimi-k3_max | Moonshot | 0.23 |
| 13 | glm-5.2_max | Z.ai (Zhipu AI) | 0.21 |
| 14 | claude-opus-4-8_max | Anthropic | 0.21 |
| 15 | gpt-5.6-luna_max | OpenAI | 0.21 |
| 16 | gpt-5.5_medium | OpenAI | 0.19 |
| 17 | gemini-3.1-pro-preview | Google DeepMind | 0.18 |
| 18 | claude-sonnet-5_max | Anthropic | 0.17 |
| 19 | muse-spark-1.1 | Meta AI | 0.15 |
| 20 | qwen3.7-max | Alibaba | 0.13 |
| 21 | gemini-3.5-flash_high | 0.13 | |
| 22 | deepseek-v4-pro_max | DeepSeek | 0.13 |
| 23 | gpt-5-2025-08-07_high | OpenAI | 0.13 |
| 24 | claude-opus-4-7_max | Anthropic | 0.12 |
| 25 | muse-spark | Meta AI | 0.11 |
| 26 | gpt-5.4-mini-2026-03-17_xhigh | OpenAI | 0.1 |
| 27 | kimi-k2.7-code | Moonshot | 0.1 |
| 28 | gpt-5.4-nano-2026-03-17_xhigh | OpenAI | 0.09 |
| 29 | gpt-5.5_low | OpenAI | 0.08 |
| 30 | kimi-k2.6 | Moonshot | 0.08 |
| 31 | grok-4.3_high | xAI | 0.08 |
| 32 | deepseek-v4-flash_max | DeepSeek | 0.07 |
| 33 | gemini-3-pro-preview | Google DeepMind | 0.07 |
| 34 | gpt-5.1-2025-11-13_unknown | OpenAI | 0.05 |
| 35 | glm-5.1 | Z.ai (Zhipu AI) | 0.05 |
| 36 | mimo-v2.5-pro | Xiaomi Corp | 0.04 |
| 37 | MiniMax-M3 | MiniMax | 0.04 |
| 38 | nemotron-3-ultra | NVIDIA | 0.03 |
| 39 | claude-sonnet-4-6_max | Anthropic | 0.03 |
| 40 | qwen3.6-plus | Alibaba | 0.03 |
| 41 | gemini-2.5-pro | Google DeepMind | 0.02 |
| 42 | gpt-5.5_none | OpenAI | 0.01 |
| 43 | gemma-4-31b-it | Google DeepMind | 0.01 |
| 44 | gpt-oss-20b_high | OpenAI | 0.01 |
| 45 | o3-2025-04-16_high | OpenAI | 0.01 |
| 46 | gemini-3.1-flash-lite | 0.01 | |
| 47 | openai/gpt-oss-120b_high | OpenAI | 0.01 |
| 48 | claude-sonnet-4-5-20250929_unknown | Anthropic | 0.01 |
| 49 | gemini-2.5-flash | Google DeepMind | 0.01 |
| 50 | DeepSeek-R1 | DeepSeek | 0.01 |
| 51 | o4-mini-2025-04-16_high | OpenAI | 0.01 |
| 52 | MiniMax-M2.7 | MiniMax | 0.01 |
| 53 | claude-opus-4-20250514_unknown | Anthropic | 0 |
| 54 | qwen3.5-9B | Alibaba | 0 |
| 55 | qwen3.6-35b-a3b | Alibaba | 0 |
| 56 | claude-sonnet-4-20250514_unknown | Anthropic | 0 |
| 57 | Qwen3-32B | Alibaba | 0 |
| 58 | Llama-3.1-8B-Instruct | Meta AI | 0 |
| 59 | gpt-4.1-mini-2025-04-14 | OpenAI | 0 |
| 60 | gpt-5-2025-08-07_minimal | OpenAI | 0 |
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