OpenCompass LLM - Reasoning - Chinese (CompassBench 2501): leaderboard
Metric: Score (%). Source: rank.opencompass.org.cn. 40 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek R1 | 86.2 |
| 2 | Doubao-1.5-pro-32k-250115 | 82.3 |
| 3 | O1 (2024-12-17) | 81.5 |
| 4 | O3 Mini (2025-01-31) | 80 |
| 5 | O1 Mini (2024-09-12) | 73.2 |
| 6 | Gemini 2.0 Pro (Preview 02-05) | 72.9 |
| 7 | Gemini 2.0 Flash (001) | 72.8 |
| 8 | QwQ 32B-Preview | 72.5 |
| 9 | Qwen 2.5 Max | 70.2 |
| 10 | GLM-Zero-Preview | 70.2 |
| 11 | DeepSeek V3 | 69.2 |
| 12 | Claude 3.5 Sonnet (20241022) | 65.7 |
| 13 | Qwen 2.5 72B Instruct | 64.9 |
| 14 | GLM-4 Plus | 64.8 |
| 15 | Yi Lightning | 61.9 |
Interactive version: theaggregate.ai/benchmark?slug=opencompass-llm-reasoning-chinese-compassbench-2501 · How It Works · Data refreshed daily, snapshot 2026-09-19.