SuperCLUE General (November 2025) - Precise Instruction Following: leaderboard

Metric: Score. Source: www.superclueai.com. 32 models tracked.

Top models

#ModelScore
1GPT-5.2 (High)55.16
2Claude Opus 4.5 (Thinking)50.32
3O3 (High)48.39
4GPT-5.1 (High)47.42
5Claude Sonnet 4.5 (Thinking)46.77
6Gemini 3 Pro (Preview)45.16
7DeepSeek V3.2 Speciale44.34
8DeepSeek V3.2 (Thinking)42.58
9DeepSeek V3.2 Exp (Thinking)37.1
10Gemini 2.5 Pro33.87
11Qwen 3 Max (Preview) (Thinking)33.55
12Qwen 3 235B A22B 2507 (Thinking)32.58
13Qwen 3 Next 80B A3B (Thinking)31.94
14Kimi K2 (Thinking)30.65
15GLM-4.730.42

Interactive version: theaggregate.ai/benchmark?slug=superclue-general-november-2025-precise-instruction-following · How It Works · Data refreshed daily, snapshot 2026-09-19.