OpenCompass Knowledge - Engineering - Chinese (CompassBench 2504): leaderboard
OpenCompass CompassBench 2504 edition (file dated 2025-05-13): engineering sub-task of the knowledge 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 | Step 2 16K | 95 | #442 |
| 2 | GPT-4.1 (2025-04-14) | 94 | #257 |
| 3 | DeepSeek R1 | 93 | #262 |
| 4 | O4 Mini (High) | 93 | #185 (O4 Mini) |
| 5 | Claude 3.7 Sonnet (Thinking) | 93 | #259 (Claude 3.7 Sonnet) |
| 6 | Gemini 2.5 Pro (Preview 05-06) | 93 | #156 |
| 7 | ERNIE-X1-Turbo-32K | 93 | #288 |
| 8 | GPT-4.1 Mini | 91 | #363 |
| 9 | MiniMax-Text-01 | 91 | #519 |
| 10 | Grok 3 | 90 | #313 |
| 11 | Llama 4 Maverick Instruct | 90 | #467 |
| 12 | Llama 4 Scout Instruct | 89 | #547 |
| 13 | Qwen 3 235B A22B | 89 | #323 |
| 14 | Qwen 3 235B A22B (Thinking) | 89 | #323 (Qwen 3 235B A22B) |
| 15 | Spark X1 | 89 | #435 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=opencompass-knowledge-engineering-chinese-compassbench-2504 · How It Works · Data refreshed daily, snapshot 2026-11-05.