OpenCompass Language - Translation (CompassBench 2405): leaderboard
OpenCompass CompassBench 2405 edition (file dated 2024-06-12): multilingual translation sub-task of the language category. 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. 39 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o (2024-05-13) | 31.5 | #337 |
| 2 | GPT-4 Turbo | 29.2 | #417 |
| 3 | Qwen 1.5 110B Chat | 22.9 | #693 |
| 4 | Yi Large | 22.4 | #566 |
| 5 | GPT-3.5 Turbo | 22 | #858 |
| 6 | Qwen 1.5 72B Chat | 21.8 | #782 |
| 7 | Claude 3 Opus | 20.3 | #565 |
| 8 | Qwen 1.5 32B Chat | 18.9 | #820 |
| 9 | Command-R+ | 18.2 | #869 |
| 10 | internlm2-chat-7B | 18.2 | #1113 |
| 11 | Qwen 1.5 14B Chat | 17 | #976 |
| 12 | internlm2-chat-20B | 16.7 | #988 |
| 13 | GLM-4 | 16.7 | #647 |
| 14 | Mixtral 8x22B Instruct | 15.4 | #738 |
| 15 | Qwen 1.5 7B Chat | 15.3 | #1292 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=opencompass-language-translation-compassbench-2405 · How It Works · Data refreshed daily, snapshot 2026-11-05.