OpenCompass Knowledge - Science - English (CompassBench 2407): leaderboard
Metric: Score (%). Source: rank.opencompass.org.cn. 36 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2 72B Instruct | 92 |
| 2 | Mistral Large | 91.2 |
| 3 | Claude 3.5 Sonnet (20240620) | 90.4 |
| 4 | Mistral Large 2 (Jul) | 90.4 |
| 5 | Llama 3.1 70B Instruct | 89.6 |
| 6 | GPT-4o (2024-05-13) | 89.6 |
| 7 | GPT-4o Mini (2024-07-18) | 88.8 |
| 8 | Mixtral 8x22B Instruct (v0.1) | 85.6 |
| 9 | GLM-4 (0520) | 84.8 |
| 10 | Qwen 1.5 110B Chat | 84 |
| 11 | DeepSeek V2 Chat (0628) | 83.2 |
| 12 | Llama 3.1 8B Instruct | 82.4 |
| 13 | Yi Large | 80.8 |
| 14 | GLM-4 9B Chat | 79.2 |
| 15 | Yi 1.5 34B Chat | 76 |
Interactive version: theaggregate.ai/benchmark?slug=opencompass-knowledge-science-english-compassbench-2407 · How It Works · Data refreshed daily, snapshot 2026-09-19.