SuperCLUE-Writing - Constrained Writing: 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 Chinese writing under explicit constraints.
Metric: Score. Source: www.superclueai.com. Saturation forecast: Around August 2027. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 (xHigh) | 96.17 | #31 (GPT-5.5) |
| 2 | Claude Opus 4.7 (Max) | 95.41 | #49 (Claude Opus 4.7) |
| 3 | Gemini 3.1 Pro (Preview) | 93.94 | #58 |
| 4 | Grok 4.20 0309 (Reasoning) | 93 | #112 (Grok 4.20 0309) |
| 5 | GLM-5.1 | 92.53 | #97 |
| 6 | Qwen 3.6 Max (Preview) (Thinking) | 92.51 | |
| 7 | Doubao-Seed-2.0-Pro-260215 | 91.03 | #77 |
| 8 | Kimi K2.6 (Thinking) | 86.05 | #104 (Kimi K2.6) |
| 9 | Ernie 5.1 | 84.47 | #202 |
| 10 | MiniMax-M2.7 | 76.72 | #258 |
| 11 | DeepSeek V4 Pro (Max) | 74.62 | #102 (DeepSeek V4 Pro) |
| 12 | Hy3-preview (High) | 74.03 | #157 (Hy3-preview) |
| 13 | DeepSeek V4 Flash (Max) | 65.16 | #137 (DeepSeek V4 Flash) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=superclue-writing-constrained-writing · How It Works · Data refreshed daily, snapshot 2026-10-20.