OpenCompass Reasoning - Science and Engineering - Chinese (CompassBench 2409): leaderboard
Metric: Score (%). Source: rank.opencompass.org.cn. 30 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GLM-4 Plus | 65 |
| 2 | Step 2 16K | 61.4 |
| 3 | DeepSeek V2.5 | 60.6 |
| 4 | GPT-4o (2024-05-13) | 58 |
| 5 | Qwen 2.5 72B Instruct | 57.7 |
| 6 | Claude 3.5 Sonnet (20240620) | 57.1 |
| 7 | Gemini 1.5 Pro | 56.1 |
| 8 | Mistral Large 2 (Jul) | 49.2 |
| 9 | Yi Large | 47.5 |
| 10 | Qwen 2.5 7B Instruct | 47.2 |
| 11 | GPT-4o (2024-08-06) | 47.1 |
| 12 | GPT-4o Mini (2024-07-18) | 45 |
| 13 | Gemma 2 27B (IT) | 43.3 |
| 14 | Mistral-Small-Instruct-2409 | 41.1 |
| 15 | Gemma 2 9B (IT) | 39.7 |
Interactive version: theaggregate.ai/benchmark?slug=opencompass-reasoning-science-and-engineering-chinese-compassbench-2409 · How It Works · Data refreshed daily, snapshot 2026-09-19.