OpenCompass Subjective - Conversation (CompassBench 2405): leaderboard
OpenCompass CompassBench 2405 edition (file dated 2024-06-12): subjective open-ended evaluation, conversation score. 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) | 69.3 | #337 |
| 2 | Qwen 1.5 110B Chat | 59.4 | #693 |
| 3 | GPT-4 Turbo | 54.9 | #417 |
| 4 | GLM-4 | 54.2 | #647 |
| 5 | Yi 1.5 34B Chat | 51.6 | #798 |
| 6 | Nanbeige2-16B-Chat | 51 | #971 |
| 7 | DeepSeek V2 Chat | 48.1 | #572 |
| 8 | Yi Large | 43.9 | #566 |
| 9 | internlm2-chat-20B | 41.7 | #988 |
| 10 | internlm2-chat-7B | 40.9 | #1113 |
| 11 | Moonshot v1 8K | 40.7 | #622 |
| 12 | Claude 3 Opus | 40.6 | #565 |
| 13 | Command-R+ | 40.3 | #869 |
| 14 | Qwen 1.5 72B Chat | 38.4 | #782 |
| 15 | Qwen 1.5 14B Chat | 38.3 | #976 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=opencompass-subjective-conversation-compassbench-2405 · How It Works · Data refreshed daily, snapshot 2026-11-05.