SuperCLUE-CCSWE - Overall: 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-context software engineering.
Metric: Score. Source: www.superclueai.com. Saturation forecast: Around November 2027. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Kimi K3 (Max) | 75.79 | #30 (Kimi K3) |
| 2 | GLM-5.2 (Max) | 64.13 | #73 (GLM-5.2) |
| 3 | DeepSeek V4 Flash (0731) (Max) | 58.95 | #78 (DeepSeek V4 Flash (0731)) |
| 4 | Kimi K2.7 Code | 55.43 | #89 |
| 5 | Doubao-Seed-2.1-Pro (High) | 51.58 | |
| 6 | MiMo-V2.5-Pro | 49.47 | #158 |
| 7 | Qwen 3.8 Max (xHigh) | 49.47 | #35 (Qwen 3.8 Max) |
| 8 | Qwen 3.7 Max (Thinking) | 48.42 | #75 (Qwen 3.7 Max) |
| 9 | DeepSeek V4 Pro (Max) | 47.37 | #102 (DeepSeek V4 Pro) |
| 10 | Hy3 (High) | 47.37 | #93 (Hy3) |
| 11 | MiniMax-M3 | 46.32 | #136 |
| 12 | LongCat 2.0 | 42.11 | #219 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=superclue-ccswe-overall · How It Works · Data refreshed daily, snapshot 2026-10-20.