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
1o1Gemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH95.3%
2QwQ 32B PreviewGemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH95.3%
3Gemini 2.0 Flash ExperimentalGemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct, thinking· U-MATH and mu-MATH93%
4QwQ 32B PreviewQwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH92.1%
5Gemini 2.0 Flash ExperimentalGPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct, thinking· U-MATH and mu-MATH91.7%
6QwQ 32B Previewmu_math, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH91.3%
7o1mu_math, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH90.6%
8o1Llama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH90.5%
9QwQ 32B PreviewLlama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH90.5%
10Gemini 2.0 Flash Experimentalmu_math, TPR, Qwen/Qwen2.5-72B-Instruct, thinking· U-MATH and mu-MATH89.1%
11o1GPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH88.9%
12O1 MiniGemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH88.4%
13Gemini 2.0 Flash ExperimentalQwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct, thinking· U-MATH and mu-MATH86.8%
14o1Qwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH86.8%
15QwQ 32B PreviewGPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH86.1%
16DeepSeek R1Llama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH85.7%
17O1 MiniLlama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH85.7%
18Qwen2.5 Math 7B InstructLlama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH85.6%
19O1 Minimu_math, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH83.3%
20Qwen2.5 Math 72B InstructGemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH82.5%
21O1 MiniQwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH81.6%
22DeepSeek R1Gemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH81.4%
23Qwen2.5 7B InstructGPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH81.4%
24Qwen2.5 Math 72B InstructLlama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH81.4%
25Qwen2.5 Math 72B InstructQwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH81.3%
26Gemini 1.5 ProGemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH81%
27Gemini 2.0 Flash ExperimentalLlama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct, thinking· U-MATH and mu-MATH81%
28Qwen2.5 Math 72B Instructmu_math, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH80.9%
29Mistral Large Instruct 2411Gemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH79.9%
30Qwen2.5 72B InstructGemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH79.4%
31Qwen2.5 7B InstructGemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH79.4%
32Qwen2.5 7B Instructmu_math, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH78.7%
33Qwen2.5 7B InstructLlama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH78.4%
34Qwen2.5 Math 72B InstructGPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH77.9%
35O1 MiniGPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH77.8%
36Qwen2.5 Math 7B InstructGemini-1.5-Pro, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH77.8%
37Gemini 1.5 Promu_math, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH77.5%
38Qwen2.5 72B Instructmu_math, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH77.1%
39DeepSeek R1mu_math, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH76.8%
40Qwen2.5 Math 7B Instructmu_math, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH76.6%
41Gemini 1.5 ProGPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH76.4%
42Qwen2.5 72B InstructGPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH76.4%
43Qwen2.5 72B InstructQwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH76.1%
44Mistral Large Instruct 2411GPT-4o, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH75.7%
45Mistral Large Instruct 2411mu_math, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH75.7%
46Gemini 1.5 ProQwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH75.5%
47Qwen2.5 7B InstructQwen2.5-72B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH75.5%
48Gemini 1.5 ProLlama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH75.3%
49Mistral Large Instruct 2411Llama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH75.3%
50Qwen2.5 72B InstructLlama-3.1-70B-Instruct, TPR, Qwen/Qwen2.5-72B-Instruct· U-MATH and mu-MATH75.3%

Showing 1–50 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)