SuperCLUE General (March 2025) - Science Reasoning: leaderboard
Metric: Score. Source: www.superclueai.com. 36 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.0 Pro (Preview 02-05) | 70.71 |
| 2 | O3 Mini (High) | 70 |
| 3 | GPT-4.5 (Preview) | 70 |
| 4 | Doubao-1.5-pro-32k-250115 | 70 |
| 5 | QwQ-32B | 67 |
| 6 | DeepSeek R1 | 64 |
| 7 | GLM-Zero-Preview | 64 |
| 8 | DeepSeek R1 Distill Qwen 14B | 63.27 |
| 9 | DeepSeek V3 | 63 |
| 10 | Gemini 2.0 Flash (01-21) (Thinking) | 63 |
| 11 | hunyuan-turbos-20250226 | 63 |
| 12 | DeepSeek R1 Distill Qwen 32B | 62.89 |
| 13 | Phi-4 | 61 |
| 14 | Claude 3.7 Sonnet | 59 |
| 15 | DeepSeek-R1-Distill-Qwen-7B | 58.06 |
Interactive version: theaggregate.ai/benchmark?slug=superclue-general-march-2025-science-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-19.