SuperCLUE General (May 2025) - Math Reasoning: leaderboard
Metric: Score. Source: www.superclueai.com. 40 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek R1 | 62.1 |
| 2 | Doubao-1.5-thinking-pro-250415 | 62.1 |
| 3 | SenseNova V6 Reasoner | 62.1 |
| 4 | DeepSeek R1 Distill Qwen 32B | 60.33 |
| 5 | Kimi K1.5 | 59.68 |
| 6 | O4 Mini (High) | 57.72 |
| 7 | ERNIE-X1-Turbo-32K | 57.63 |
| 8 | Gemini 2.5 Flash (Preview 04-17) | 57.26 |
| 9 | O3 | 56.45 |
| 10 | Qwen 3 32B (Thinking) | 56.45 |
| 11 | Gemini 2.5 Flash (Preview 05-20) | 56.45 |
| 12 | DeepSeek V3 (0324) | 55.65 |
| 13 | Gemini 2.5 Pro (Preview 05-06) | 55.65 |
| 14 | Claude Opus 4 (Thinking) | 51.61 |
| 15 | Llama 4 Maverick Instruct | 44.35 |
Interactive version: theaggregate.ai/benchmark?slug=superclue-general-may-2025-math-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-19.