OpenCompass LLM - Math - Chinese (CompassBench 2501): leaderboard
Metric: Score (%). Source: rank.opencompass.org.cn. 40 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | O3 Mini (2025-01-31) | 65.2 |
| 2 | O1 (2024-12-17) | 65 |
| 3 | DeepSeek R1 | 61.4 |
| 4 | O1 Mini (2024-09-12) | 51.1 |
| 5 | GLM-Zero-Preview | 42 |
| 6 | Gemini 2.0 Pro (Preview 02-05) | 40.6 |
| 7 | QwQ 32B-Preview | 40.3 |
| 8 | Doubao-1.5-pro-32k-250115 | 38.4 |
| 9 | Gemini 2.0 Flash (001) | 35.7 |
| 10 | DeepSeek V3 | 33.1 |
| 11 | Qwen 2.5 Max | 26.6 |
| 12 | GPT-4o (2024-11-20) | 26.4 |
| 13 | Mistral Large 2 (Nov) Instruct (2411) | 24 |
| 14 | Qwen 2.5 72B Instruct | 23.6 |
| 15 | Qwen 2.5 7B Instruct | 22.2 |
Interactive version: theaggregate.ai/benchmark?slug=opencompass-llm-math-chinese-compassbench-2501 · How It Works · Data refreshed daily, snapshot 2026-09-19.