Skip to content

Benchmark

U-MATH differential_calc accuracy

Math benchmarks measure how well models solve maths problems. Scores are percentages. Higher is better.

Models measured
30
Scores recorded
90
Sources
1
IQ weight
2.13

Score distribution

35.2%
Mean
30.7%
Median
20%
Std dev
11.4%
10th percentile
65.5%
90th percentile
4.3%
Min
4.3%88%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.13 for chat models, measured over 30 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51Qwen2-VL (72B) Instructdifferential_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH29.3%
52GPT-4o minidifferential_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH28.2%
53Llama 3.3 70B Instructdifferential_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH27.7%
54Claude 3.5 Sonnetdifferential_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH27.3%
55Llama 3.2 90B Vision Instructdifferential_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH27.1%
56Nvidia Llama 3.1 Nemotron 70B Instructdifferential_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH26.8%
57GPT-4o minidifferential_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH26%
58Mistral Large Instruct 2411differential_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH24.5%
59Qwen2 VL 7b Instructdifferential_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH24.5%
60Qwen2.5 Math 7B Instructdifferential_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH24.5%
61LFM-7bdifferential_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH24.3%
62QwQ 32B Previewdifferential_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH24.3%
63Llama 3.2 90B Vision Instructdifferential_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH23.2%
64DeepSeek R1differential_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH22.9%
65Qwen2.5 7B Instructdifferential_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH22.3%
66Nvidia Llama 3.1 Nemotron 70B Instructdifferential_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH21.4%
67Llama 3.2 90B Vision Instructdifferential_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH21.3%
68Claude 3.5 Sonnetdifferential_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH20.7%
69Llama 3.2 11B Vision Instructdifferential_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH20%
70Qwen2 VL 7b Instructdifferential_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH18.7%
71Llama 3.1 8B Instructdifferential_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH17.3%
72o3 Minidifferential_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH17.1%
73Pixtral 12Bdifferential_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH16.8%
74Llama 3.1 70B Instructdifferential_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH15.7%
75Qwen2.5 72B Instructdifferential_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH15.7%
76Llama 3.1 70B Instructdifferential_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH15.5%
77Llama 3.1 70B Instructdifferential_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH15.3%
78Llama 3.1 8B Instructdifferential_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH15%
79LFM-7bdifferential_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH13.6%
80Qwen2.5 Math 72B Instructdifferential_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH12.9%
81Llama 3.3 70B Instructdifferential_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH11.4%
82Qwen2.5 32B Instructdifferential_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH11.4%
83Llama 3.2 11B Vision Instructdifferential_calc, all, gpt-4o-2024-08-06· U-MATH and mu-MATH11.4%
84Pixtral 12Bdifferential_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH10.7%
85Llama 3.1 8B Instructdifferential_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH10%
86LFM-7bdifferential_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH8.7%
87Mistral Large Instruct 2411differential_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH8.6%
88Qwen2.5 Math 7B Instructdifferential_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH8.6%
89Llama 3.2 11B Vision Instructdifferential_calc, text, gpt-4o-2024-08-06· U-MATH and mu-MATH7.3%
90Qwen2.5 7B Instructdifferential_calc, visual, gpt-4o-2024-08-06· U-MATH and mu-MATH4.3%

Showing 51–90 of 90 results.

Data credits: Toloka, U-MATH and mu-MATH public leaderboard. MIT Space; generated solution outputs retain underlying model terms and are not ingested. (licence)