Skip to content

Benchmark

mu-MATH TPR

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

Models measured
17
Scores recorded
85
Sources
1
IQ weight
0.53

Score distribution

74.2%
Mean
76.1%
Median
12%
Std dev
57.9%
10th percentile
90.5%
90th percentile
45.2%
Min
45.2%95.3%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.53 for chat models, measured over 17 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
51Qwen2.5 Math 7B InstructGPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH75%
52DeepSeek R1GPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH72.2%
53GPT 4oGPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH72.1%
54DeepSeek R1Qwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH71.1%
55GPT 4oLlama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH71.1%
56Mistral Large Instruct 2411Qwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH71%
57Qwen2.5 Math 7B InstructQwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH71%
58Claude 3.5 SonnetGemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH70.9%
59GPT 4oGemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH70.4%
60GPT 4omu_math, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH70.1%
61Gemini 1.5 FlashGemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH67.7%
62GPT 4oQwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH67.1%
63Llama 3.1 70B InstructGPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH67.1%
64Gemini 1.5 FlashLlama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH67%
65Llama 3.1 70B InstructLlama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH63.9%
66Gemini 1.5 Flashmu_math, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH63.3%
67GPT-4o miniGemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH63%
68Claude 3.5 Sonnetmu_math, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH62.5%
69Llama 3.1 70B Instructmu_math, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH62.5%
70Llama 3.1 70B InstructGemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH61.4%
71Gemini 1.5 FlashQwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH60.6%
72Claude 3.5 SonnetLlama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH59.8%
73GPT-4o miniLlama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH59.8%
74GPT-4o minimu_math, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH59%
75Llama 3.1 70B InstructQwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH58.7%
76Claude 3.5 SonnetQwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH58.1%
77Claude 3.5 SonnetGPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH57.9%
78Gemini 1.5 FlashGPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH57.9%
79GPT-4o miniGPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH56.4%
80GPT-4o miniQwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH56.1%
81Llama 3.1 8B InstructGemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH55%
82Llama 3.1 8B Instructmu_math, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH48.7%
83Llama 3.1 8B InstructGPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH46.4%
84Llama 3.1 8B InstructLlama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH45.4%
85Llama 3.1 8B InstructQwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH45.2%

Showing 51–85 of 85 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)