OpenCompass LLM - Math - Chinese (CompassBench 2407): leaderboard
Metric: Score (%). Source: rank.opencompass.org.cn. 36 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4o (2024-05-13) | 76.2 |
| 2 | Claude 3.5 Sonnet (20240620) | 74.7 |
| 3 | Mistral Large 2 (Jul) | 74.5 |
| 4 | DeepSeek V2 Chat (0628) | 72 |
| 5 | Mistral Large | 71.3 |
| 6 | GPT-4o Mini (2024-07-18) | 68.7 |
| 7 | GLM-4 (0520) | 61.5 |
| 8 | Yi Large | 60.8 |
| 9 | Gemma 2 27B (IT) | 60.1 |
| 10 | Llama 3.1 70B Instruct | 59.7 |
| 11 | Qwen 2 72B Instruct | 58.2 |
| 12 | Moonshot v1 8K | 56.2 |
| 13 | Mixtral 8x22B Instruct (v0.1) | 52.4 |
| 14 | Gemma 2 9B (IT) | 49.3 |
| 15 | Llama 3.1 8B Instruct | 45.7 |
Interactive version: theaggregate.ai/benchmark?slug=opencompass-llm-math-chinese-compassbench-2407 · How It Works · Data refreshed daily, snapshot 2026-09-19.