OpenCompass Code - Competition - Chinese (CompassBench 2504): leaderboard
OpenCompass CompassBench 2504 edition (file dated 2025-05-13): competition programming sub-task of the code category, Chinese questions. 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. 28 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O4 Mini (High) | 83 | #185 (O4 Mini) |
| 2 | Qwen 3 235B A22B (Thinking) | 76.6 | #323 (Qwen 3 235B A22B) |
| 3 | Qwen 3 32B (Thinking) | 74.5 | #449 (Qwen 3 32B) |
| 4 | O3 Mini (High) | 74.5 | #282 (O3 Mini) |
| 5 | DeepSeek R1 | 72.3 | #262 |
| 6 | Doubao-1.5-thinking-pro-250415 | 72.3 | #226 |
| 7 | Gemini 2.5 Pro (Preview 05-06) | 68.1 | #156 |
| 8 | ERNIE-X1-Turbo-32K | 66 | #288 |
| 9 | GLM-Z1-32B-0414 | 63.8 | #405 |
| 10 | Spark X1 | 59.6 | #435 |
| 11 | DeepSeek V3 (0324) | 55.3 | #349 |
| 12 | GPT-4.1 (2025-04-14) | 53.2 | #257 |
| 13 | GPT-4.1 Mini | 48.9 | #363 |
| 14 | Grok 3 | 48.9 | #313 |
| 15 | Claude 3.7 Sonnet (Thinking) | 48.9 | #259 (Claude 3.7 Sonnet) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=opencompass-code-competition-chinese-compassbench-2504 · How It Works · Data refreshed daily, snapshot 2026-11-05.