OpenCompass Math - College - Chinese (CompassBench 2504): leaderboard

Metric: Score (%). Source: rank.opencompass.org.cn. 28 models tracked.

Top models

#ModelScore
1DeepSeek R160
2O3 Mini (High)56
3Qwen 3 32B (Thinking)56
4Gemini 2.5 Pro (Preview 05-06)56
5Grok 352
6DeepSeek V3 (0324)52
7O4 Mini (High)52
8Qwen 3 235B A22B (Thinking)52
9Doubao-1.5-thinking-pro-25041552
10GPT-4.1 Mini48
11GPT-4.1 (2025-04-14)48
12Qwen 3 235B A22B48
13GLM-4 32B (0414)48
14Claude 3.7 Sonnet (Thinking)48
15Llama 4 Scout Instruct44

Interactive version: theaggregate.ai/benchmark?slug=opencompass-math-college-chinese-compassbench-2504 · How It Works · Data refreshed daily, snapshot 2026-09-19.