SuperCLUE-SWE - 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 software engineering.
Metric: Score. Source: www.superclueai.com. Saturation forecast: Around September 2027. 17 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.8 (High) | 73.33 | #37 (Claude Opus 4.8) |
| 2 | Gemini 3.1 Pro (Preview) (High) | 66.67 | #58 (Gemini 3.1 Pro (Preview)) |
| 3 | Qwen 3.7 Max (Thinking) | 66.67 | #75 (Qwen 3.7 Max) |
| 4 | GLM-5.2 (Max) | 63.33 | #73 (GLM-5.2) |
| 5 | Claude Opus 4.7 (High) | 60 | #49 (Claude Opus 4.7) |
| 6 | Kimi K2.7 Code | 56.67 | #89 |
| 7 | DeepSeek V4 Pro (Max) | 53.33 | #102 (DeepSeek V4 Pro) |
| 8 | Kimi K2.6 (Thinking) | 53.33 | #104 (Kimi K2.6) |
| 9 | Gemini 3.5 Flash (High) | 50 | #60 (Gemini 3.5 Flash) |
| 10 | Doubao-Seed-2-0-Pro-260215 (High) | 50 | |
| 11 | GPT-5.5 (High) | 43.33 | #31 (GPT-5.5) |
| 12 | GLM-5.1 | 40 | #97 |
| 13 | Step 3.7 Flash (High) | 40 | #236 (Step 3.7 Flash) |
| 14 | Grok 4.3 | 36.67 | #188 |
| 15 | MiMo-V2.5-Pro | 33.33 | #158 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=superclue-swe-overall · How It Works · Data refreshed daily, snapshot 2026-10-20.