OpenCompass LLM - Math (CompassBench 2501): leaderboard
OpenCompass CompassBench 2501 edition (file dated 2025-02-20): math 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. 40 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O3 Mini (2025-01-31) | 70.7 | #249 |
| 2 | O1 (2024-12-17) | 70.5 | #157 |
| 3 | DeepSeek R1 | 63.1 | #262 |
| 4 | O1 Mini (2024-09-12) | 53 | #412 |
| 5 | QwQ 32B-Preview | 43.2 | #671 |
| 6 | GLM-Zero-Preview | 42 | #682 |
| 7 | DeepSeek V3 | 40.6 | #329 |
| 8 | Doubao-1.5-pro-32k-250115 | 40 | #320 |
| 9 | Gemini 2.0 Pro (Preview 02-05) | 38.7 | #318 |
| 10 | Gemini 2.0 Flash (001) | 37.6 | #365 |
| 11 | Qwen 2.5 Max | 29.4 | #427 |
| 12 | Qwen 2.5 72B Instruct | 26.9 | #469 |
| 13 | MiniMax-Text-01 | 24.9 | #519 |
| 14 | Claude 3.5 Sonnet (20241022) | 24.2 | #302 |
| 15 | Qwen 2.5 7B Instruct | 23.6 | #863 |
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-math-compassbench-2501 · How It Works · Data refreshed daily, snapshot 2026-11-05.