SuperCLUE General (May 2026) - Agentic Task Planning: leaderboard

Metric: Score. Source: www.superclueai.com. 24 models tracked.

Top models

#ModelScore
1GPT-5.5 (High)86.56
2Qwen 3.6 Max (Preview) (Thinking)83.41
3Kimi K2.6 (Thinking)80.95
4DeepSeek V4 Pro (Max)78.12
5MiniMax-M375.91
6Claude Opus 4.7 (High)75.7
7DeepSeek V4 Flash (Max)75.56
8Gemini 3.1 Pro (Preview) (High)75.12
9Spark X271.83
10Claude Opus 4.8 (High)71.63
11Ernie 5.170.44
12Gemini 3.5 Flash (High)69.6
13Step 3.7 Flash67.24
14GLM-5.166.8
15Step 3.5 Flash64.99

Interactive version: theaggregate.ai/benchmark?slug=superclue-general-may-2026-agentic-task-planning · How It Works · Data refreshed daily, snapshot 2026-09-19.