OpenCompass Reasoning - Common Sense - English (CompassBench 2409): leaderboard

Metric: Score (%). Source: rank.opencompass.org.cn. 30 models tracked.

Top models

#ModelScore
1GLM-4 Plus58.2
2Step 2 16K57.8
3Claude 3.5 Sonnet (20240620)54
4Qwen 2.5 72B Instruct53.8
5DeepSeek V2.550
6GPT-4o (2024-05-13)49.8
7Mistral Large 2 (Jul)49.2
8Mistral Nemo Instruct (2407)48.8
9Llama 3.1 405B Instruct FP846.8
10Gemini 1.5 Pro46
11Yi-1.5-9B Chat45
12Mistral-Small-Instruct-240945
13Gemma 2 27B (IT)44.8
14Qwen 2.5 7B Instruct44.8
15Yi Large44

Interactive version: theaggregate.ai/benchmark?slug=opencompass-reasoning-common-sense-english-compassbench-2409 · How It Works · Data refreshed daily, snapshot 2026-09-19.