OpenCompass Language - NLP — leaderboard

Metric: Score (%). Source: rank.opencompass.org.cn. 23 models tracked.

Top models

#ModelScore
1GPT-5.4 (High)74.6
2Qwen 3.6 Max Preview72.9
3GPT-5.4 Mini (High)69.6
4Gemini 3.1 Pro (Preview)67.5
5Kimi K2.667.5
6Claude Sonnet 4.6 (High)66.2
7Claude Opus 4.7 (High)65.4
8GLM-5.163.8
9DeepSeek V4 Pro63.3
10MiniMax-M2.763.3
11Qwen 3.6 27B63.3
12Hy3-preview (High)63.3
13Qwen 3.5 397B A17B62.5
14DeepSeek V4 Flash62.5
15JT-MINI62.1

Interactive version: theaggregate.ai/benchmark?slug=opencompass-language-nlp · How the rankings work · Data refreshed daily, snapshot 2026-07-22.