U-MATH - Algebra — leaderboard

Metric: Accuracy (%). Source: huggingface.co. 28 models tracked.

Top models

#ModelScore
1O189.44
2Gemini 2.0 Flash (01-21) (Thinking)89.44
3O1 Mini87.78
4Gemini 1.5 Pro86.11
5O3 Mini84.44
6QVQ-72B-Preview83.89
7Gemini 1.5 Flash83.33
8DeepSeek R183.33
9Phi-481.67
10DeepSeek V381.67
11GPT-4o (2024-08-06)80.56
12QwQ 32B-Preview80
13Qwen2.5-Math-72B-Instruct80
14Qwen 2.5 32B Instruct78.89
15Qwen 2.5 72B Instruct78.33

Interactive version: aibenchmarks.dev/benchmark?slug=u-math-algebra · How the rankings work · Data refreshed daily, snapshot 2026-07-08.