OpenCompass LLM - Reasoning (CompassBench 2407): leaderboard
OpenCompass CompassBench 2407 edition (file dated 2024-08-05): reasoning category score. CompassBench is OpenCompass's periodic evaluation of chat models on a private question set; each edition is a separate board. Score 0-100, higher is better.
Metric: Score (%). Source: rank.opencompass.org.cn. Saturation forecast: Estimated already saturated. 36 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude 3.5 Sonnet (20240620) | 57 | #308 |
| 2 | GPT-4o (2024-05-13) | 55.8 | #337 |
| 3 | Mistral Large | 50.1 | #713 |
| 4 | Mistral Large 2 (Jul) | 50 | #490 |
| 5 | Yi Large | 47.6 | #566 |
| 6 | DeepSeek V2 Chat (0628) | 47.4 | #409 |
| 7 | GLM-4 (0520) | 46.1 | #612 |
| 8 | Moonshot v1 8K | 46 | #622 |
| 9 | Qwen 1.5 110B Chat | 45.8 | #693 |
| 10 | GPT-4o Mini (2024-07-18) | 45.4 | #568 |
| 11 | Gemma 2 27B (IT) | 45.4 | #652 |
| 12 | Qwen 2 72B Instruct | 44.7 | #579 |
| 13 | Yi 1.5 34B Chat | 42.7 | #798 |
| 14 | Gemma 2 9B (IT) | 41.9 | #795 |
| 15 | Nanbeige2-16B-Chat | 40.5 | #971 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=opencompass-llm-reasoning-compassbench-2407 · How It Works · Data refreshed daily, snapshot 2026-11-05.