U-MATH - Differential Calculus — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 28 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | O1 | 76.82 |
| 2 | Gemini 2.0 Flash (01-21) (Thinking) | 70.45 |
| 3 | O1 Mini | 68.18 |
| 4 | DeepSeek R1 | 65.45 |
| 5 | O3 Mini | 65.45 |
| 6 | QwQ 32B-Preview | 55.45 |
| 7 | QVQ-72B-Preview | 50 |
| 8 | Gemini 1.5 Pro | 49.55 |
| 9 | Gemini 1.5 Flash | 47.27 |
| 10 | DeepSeek V3 | 43.18 |
| 11 | Phi-4 | 42.27 |
| 12 | Athene-V2-Chat | 36.82 |
| 13 | Qwen2.5-Math-72B-Instruct | 35.45 |
| 14 | Qwen 2.5 32B Instruct | 32.73 |
| 15 | GPT-4o (2024-08-06) | 32.27 |
Interactive version: aibenchmarks.dev/benchmark?slug=u-math-differential-calculus · How the rankings work · Data refreshed daily, snapshot 2026-07-08.