Skip to content

Benchmark

MATH Level 5

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

Models measured
221
Scores recorded
238
Sources
2
IQ weight
1.72

Score distribution

32.6%
Mean
22%
Median
30.4%
Std dev
2.3%
10th percentile
87.1%
90th percentile
0%
Min
0%98.1%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.72 for chat models, measured over 221 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 32B Instruct· Hugging Face Open LLM Leaderboard data
Values from 2 sources
62.5%
52Qwen2.5 Math 72B Instruct· Hugging Face Open LLM Leaderboard data62.4%
53Llama 4 Scout· Epoch AI Benchmarking Hub data62.3%
54gemini-1.5-flash-002· Epoch AI Benchmarking Hub data61.9%
55Qwen2.5 Math 7B Instruct· Hugging Face Open LLM Leaderboard data58.1%
56Claude 3.5 Sonnet· Epoch AI Benchmarking Hub data56.9%
57Qwen Turbo· Epoch AI Benchmarking Hub data56.2%
58Qwen 2.5 14B Instruct· Hugging Face Open LLM Leaderboard data54.8%
59GPT-4o mini· Epoch AI Benchmarking Hub data52.6%
60Mistral Large2411· Epoch AI Benchmarking Hub data50.3%
61Qwen2.5 7B Instruct· Hugging Face Open LLM Leaderboard data50%
62GPT 4o· Epoch AI Benchmarking Hub data49.8%
63Llama 3.1 405B Instruct· Epoch AI Benchmarking Hub data49.8%
64Mistral Large Instruct 2411· Hugging Face Open LLM Leaderboard data49.5%
65Qwen 2.5 Coder 32B Instruct· Hugging Face Open LLM Leaderboard data49.5%
66Llama 3.3 70B Instruct· Hugging Face Open LLM Leaderboard data
Values from 2 sources
48.3%
67Mistral Small2503· Epoch AI Benchmarking Hub data46.8%
68GPT-4 Turbo· Epoch AI Benchmarking Hub data46.7%
69Claude Haiku 3.5· Epoch AI Benchmarking Hub data46.4%
70L3.3 Nevoria R1 70b· Hugging Face Open LLM Leaderboard data46.3%
71QwQ 32B Preview· Hugging Face Open LLM Leaderboard data44.9%
72Mistral Large2407· Epoch AI Benchmarking Hub data44.8%
73Mistral Small· Epoch AI Benchmarking Hub data44.8%
74EVA-Qwen2.5-72B-v0.2· Hugging Face Open LLM Leaderboard data43.1%
75Llama 3.1 Nemotron 70B Instruct HF· Hugging Face Open LLM Leaderboard data42.7%
76Qwen2 72B Instruct· Hugging Face Open LLM Leaderboard data
Values from 2 sources
41.8%
77gemini-1.5-pro-001· Epoch AI Benchmarking Hub data40.7%
78GPT 4 1106 Preview· Epoch AI Benchmarking Hub data40%
79Qwen2 72B Dracarys· Hugging Face Open LLM Leaderboard data39.7%
80L3.3 Ms Nevoria 70b· Hugging Face Open LLM Leaderboard data39.6%
81Llama 3.2 90B Vision Instruct· Epoch AI Benchmarking Hub data39.4%
82Qwen2.5 72B· Hugging Face Open LLM Leaderboard data39.1%
83Llama 3.1 70B Instruct· Hugging Face Open LLM Leaderboard data
Values from 2 sources
38.1%
84Claude 3 Opus· Epoch AI Benchmarking Hub data37.5%
85Qwen2.5 Coder 7B Instruct· Hugging Face Open LLM Leaderboard data37.2%
86Qwen2.5 3B Instruct· Hugging Face Open LLM Leaderboard data36.8%
87Qwen2.5 32B· Hugging Face Open LLM Leaderboard data35.6%
88Magnum V2 72B· Hugging Face Open LLM Leaderboard data35.4%
89GPT-4-0125-preview· Epoch AI Benchmarking Hub data35.4%
90arcee-blitz· Hugging Face Open LLM Leaderboard data34.8%
91Qwen2-VL (72B) Instruct· Hugging Face Open LLM Leaderboard data34.4%
92qwen2.5-coder-14b-instruct· Hugging Face Open LLM Leaderboard data32.5%
93phi-3.5-moe-instruct· Hugging Face Open LLM Leaderboard data31.2%
94Qwen 2 (72B)· Hugging Face Open LLM Leaderboard data31.1%
95Qwen 2.5 Coder 32B· Hugging Face Open LLM Leaderboard data30.9%
96DeepSeek R1 Distill Llama 70B· Hugging Face Open LLM Leaderboard data30.7%
97C4AI Command R7b 12 2024· Hugging Face Open LLM Leaderboard data29.9%
98Qwen2.5 14B· Hugging Face Open LLM Leaderboard data29%
99Qwen2 57B A14B Instruct· Hugging Face Open LLM Leaderboard data28.2%
100dolphin-2-9-2-qwen2-72b· Hugging Face Open LLM Leaderboard data28%

Showing 51–100 of 238 results.

Data credits: Epoch AI Benchmarking Hub, CC BY 4.0 (licence)