SuperCLUE-Agent - Task Planning: leaderboard
SuperCLUE is the Chinese-language evaluation programme run by CLUE; its specialised boards each cover one professional domain and are graded by the organisers rather than self-reported. This board covers agent capability: decomposing a goal into a workable plan.
Metric: Score. Source: www.superclueai.com. Saturation forecast: Estimated already saturated. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o (2024-05-13) | 75.39 | #325 |
| 2 | Qwen 2 72B Instruct | 73.3 | #576 |
| 3 | TeleChat2 | 71.02 | #589 |
| 4 | GLM-4 9B Chat | 69.7 | #916 |
| 5 | Yi-1.5-6B-Chat | 64.28 | #1259 |
| 6 | Phi-3 Mini 128K Instruct | 55.47 | #1298 |
| 7 | gemma-7B (IT) | 47.1 | #1496 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=superclue-agent-task-planning · How It Works · Data refreshed daily, snapshot 2026-10-20.