SuperCLUE General (July 2026) - Math Reasoning: leaderboard
Metric: Score. Source: www.superclueai.com. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek V4.1 Flash (Max) | 84.21 |
| 2 | DeepSeek V4 Flash (Max) | 78.95 |
| 3 | DeepSeek V4 Pro (0813) (Max) | 77.19 |
| 4 | Hy3 (High) | 77.19 |
| 5 | Qwen 3.8 Max (Max) | 77.19 |
| 6 | Qwen 3.8 Flash (Max) | 75.44 |
| 7 | DeepSeek V4 Pro (Max) | 73.68 |
| 8 | GLM-5.3 Flash (Max) | 73.68 |
| 9 | Kimi K3 (Max) | 71.93 |
| 10 | GLM-5.3 (Max) | 70.59 |
| 11 | MiniMax-M3 | 70.18 |
| 12 | GLM-5.2 (Max) | 70.18 |
| 13 | LongCat 2.0 | 70.18 |
| 14 | MiMo-V2.5-Pro | 66.67 |
| 15 | Step 3.7 Flash (High) | 64.91 |
Interactive version: theaggregate.ai/benchmark?slug=superclue-general-july-2026-math-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-19.