SuperCLUE-Terminal - 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 terminal and command-line agent work.
Metric: Score. Source: www.superclueai.com. Saturation forecast: Around November 2027. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | MiMo-V2.6-Pro | 64.65 | #42 |
| 2 | Kimi K3 (Max) | 60.61 | #30 (Kimi K3) |
| 3 | GLM-5.3 (Max) | 56.57 | #39 (GLM-5.3) |
| 4 | MiMo-V2.6-Flash | 53.54 | #70 |
| 5 | DeepSeek V4 Pro (0813) (Max) | 51.52 | #53 (DeepSeek V4 Pro (0813)) |
| 6 | DeepSeek V4.1 Flash (Max) | 50.51 | #36 (DeepSeek V4.1 Flash) |
| 7 | Qwen 3.8 Max (xHigh) | 49.49 | #35 (Qwen 3.8 Max) |
| 8 | GLM-5.2 (Max) | 48.48 | #73 (GLM-5.2) |
| 9 | GLM-5.3 Flash (Max) | 48.48 | #65 (GLM-5.3 Flash) |
| 10 | DeepSeek V4 Flash (0731) (Max) | 46.46 | #78 (DeepSeek V4 Flash (0731)) |
| 11 | MiniMax-M3 | 36.36 | #136 |
| 12 | Doubao-Seed-2.1-Pro (High) | 35.35 | |
| 13 | Hy3 (High) | 34.34 | #93 (Hy3) |
| 14 | LongCat 2.0 | 33.33 | #219 |
| 15 | MiMo-V2.5-Pro | 27.27 | #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-terminal-overall · How It Works · Data refreshed daily, snapshot 2026-10-20.