OpenCompass LLM - Instruction Following (CompassBench 2501): leaderboard

Metric: Score (%). Source: rank.opencompass.org.cn. 40 models tracked.

Top models

#ModelScore
1DeepSeek R174
2GPT-4o (2024-11-20)67
3Gemini 2.0 Pro (Preview 02-05)64.8
4Yi Lightning64
5MiniMax-Text-0163.2
6GLM-4 Plus62.9
7Gemini 2.0 Flash (001)62.5
8DeepSeek V358
9GPT-4o Mini (2024-07-18)53.5
10Qwen 2.5 Max52.7
11Mistral Large 2 (Nov) Instruct (2411)52.4
12Qwen 2.5 72B Instruct48.8
13Yi 1.5 34B Chat48.3
14Claude 3.5 Sonnet (20241022)48.2
15QwQ 32B-Preview43.5

Interactive version: theaggregate.ai/benchmark?slug=opencompass-llm-instruction-following-compassbench-2501 · How It Works · Data refreshed daily, snapshot 2026-09-19.