OpenCompass LLM - Instruction Following - English (CompassBench 2501): leaderboard
OpenCompass CompassBench 2501 edition (file dated 2025-02-20): instruction-following category score, English questions. 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. 40 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o (2024-11-20) | 68.8 | #392 |
| 2 | MiniMax-Text-01 | 67.3 | #519 |
| 3 | DeepSeek R1 | 62.7 | #262 |
| 4 | Gemini 2.0 Pro (Preview 02-05) | 57.3 | #318 |
| 5 | Yi Lightning | 57.3 | #429 |
| 6 | GLM-4 Plus | 57.1 | #398 |
| 7 | Mistral Large 2 (Nov) Instruct (2411) | 56 | #480 |
| 8 | DeepSeek V3 | 55 | #329 |
| 9 | GPT-4o Mini (2024-07-18) | 53.3 | #568 |
| 10 | Llama 3.1 70B Instruct | 51.7 | #567 |
| 11 | Gemini 2.0 Flash (001) | 50.4 | #365 |
| 12 | Qwen 2.5 Max | 48.3 | #427 |
| 13 | QwQ 32B-Preview | 46.7 | #671 |
| 14 | Llama 3.1 405B Instruct FP8 | 46 | #400 |
| 15 | Yi 1.5 34B Chat | 45.8 | #798 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=opencompass-llm-instruction-following-english-compassbench-2501 · How It Works · Data refreshed daily, snapshot 2026-11-05.