OpenCompass Agent - CIBench - Chinese (CompassBench 2402): leaderboard
OpenCompass CompassBench 2402 edition (file dated 2024-02-20): data-science code interpreter tasks (CIBench) sub-task of the agent 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. 35 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4 Turbo (Preview) | 68.5 | #416 |
| 2 | GPT-3.5 Turbo | 66.9 | #858 |
| 3 | MiniMax-abab5.5 | 65.6 | #1030 |
| 4 | GLM-4 | 58.5 | #647 |
| 5 | Qwen 1.5 14B Chat | 57.5 | #976 |
| 6 | internlm2-chat-20B | 53.9 | #988 |
| 7 | Mixtral 8x7B Instruct | 49.6 | #1085 |
| 8 | Qwen Max | 46.2 | #388 |
| 9 | Qwen-14B-Chat | 44.1 | #1161 |
| 10 | internlm2-chat-7B | 43.6 | #1113 |
| 11 | Qwen 1.5 4B Chat | 41.6 | #1442 |
| 12 | Qwen-72B-Chat | 38.8 | #835 |
| 13 | Yi Chat 34B | 38.2 | #950 |
| 14 | Qwen 1.5 7B Chat | 37.5 | #1292 |
| 15 | Mistral 7B Instruct (v0.2) | 36.8 | #1311 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=opencompass-agent-cibench-chinese-compassbench-2402 · How It Works · Data refreshed daily, snapshot 2026-11-05.