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
101Gemma 2 27B IT· Epoch AI Benchmarking Hub data
Values from 2 sources
27.9%
102Yi 1.5 34B Chat· Hugging Face Open LLM Leaderboard data
Values from 2 sources
27.7%
103Qwen2 7B Instruct· Hugging Face Open LLM Leaderboard data27.6%
104Qwen2.5 Math 1.5b Instruct· Hugging Face Open LLM Leaderboard data26.3%
105DeepHermes 3 Mistral 24B Preview· Hugging Face Open LLM Leaderboard data25.8%
106WizardLM 2 8x22B· Epoch AI Benchmarking Hub data25.7%
107internlm2-5-7b-chat· Hugging Face Open LLM Leaderboard data25.3%
108gemini-1.5-flash-001· Epoch AI Benchmarking Hub data25.1%
109Qwen2.5 7B· Hugging Face Open LLM Leaderboard data25.1%
110WizardLM 2 8x22B· Hugging Face Open LLM Leaderboard data25%
111Llama 3 70B Instruct· Hugging Face Open LLM Leaderboard data
Values from 2 sources
24.5%
112Mistral Large2402· Epoch AI Benchmarking Hub data24.5%
113Granite 3.2 8b Instruct· Hugging Face Open LLM Leaderboard data23.8%
114Qwen1.5-110B-Chat· Hugging Face Open LLM Leaderboard data23.4%
115GPT 4· Epoch AI Benchmarking Hub data23%
116Llama 3.1 8B Instruct· Epoch AI Benchmarking Hub data
Values from 2 sources
22.9%
117qwen2p5-coder-14b· Hugging Face Open LLM Leaderboard data22.5%
118Qwen2.5 1.5B Instruct· Hugging Face Open LLM Leaderboard data22.1%
119DeepSeek R1 Distill Llama 8B· Hugging Face Open LLM Leaderboard data22%
120Granite 3.1 8B Instruct· Hugging Face Open LLM Leaderboard data22%
121l3-70b-euryale-v2.1· Hugging Face Open LLM Leaderboard data21.4%
122Yi 1.5 34B Chat16k· Hugging Face Open LLM Leaderboard data21.4%
123Gemma 2 9B IT· Epoch AI Benchmarking Hub data
Values from 2 sources
21%
124Hermes 3 Llama 3.1 70B· Hugging Face Open LLM Leaderboard data21%
125Qwen 2 (7B)· Hugging Face Open LLM Leaderboard data20.4%
126Qwen1.5-14B· Hugging Face Open LLM Leaderboard data20.2%
127Qwen2 VL 7b Instruct· Hugging Face Open LLM Leaderboard data19.9%
128DeepSeek R1 Distill Qwen 7B· Hugging Face Open LLM Leaderboard data19.6%
129Phi-3 Medium 4K Instruct· Hugging Face Open LLM Leaderboard data19.6%
130Qwen1.5-32B-Chat· Hugging Face Open LLM Leaderboard data19.6%
131Phi-3 Medium 128K Instruct· Hugging Face Open LLM Leaderboard data
Values from 2 sources
19.2%
132Qwen2.5-Coder-7B· Hugging Face Open LLM Leaderboard data19.2%
133Phi-3 Small 8K Instruct· Hugging Face Open LLM Leaderboard data18.9%
134Mixtral 8x22B Instruct v0.1· Hugging Face Open LLM Leaderboard data18.7%
135Meta Llama 3 70b· Hugging Face Open LLM Leaderboard data18.6%
136Llama 3.1 70B· Hugging Face Open LLM Leaderboard data18.4%
137mixtral-8x22b-v0.1· Hugging Face Open LLM Leaderboard data18.4%
138Claude 3 Sonnet· Epoch AI Benchmarking Hub data18.2%
139Llama 3.2 3B· Hugging Face Open LLM Leaderboard data17.7%
140DeepSeek R1 Distill Qwen 32B· Hugging Face Open LLM Leaderboard data17.1%
141Phi 4 Mini· Hugging Face Open LLM Leaderboard data17%
142DeepSeek R1 Distill Qwen 1.5B· Hugging Face Open LLM Leaderboard data16.9%
143gemma2-27b· Hugging Face Open LLM Leaderboard data16.6%
144QwQ 32B· Hugging Face Open LLM Leaderboard data16.1%
145Aya Expanse 32B· Hugging Face Open LLM Leaderboard data15.3%
146Granite 3.1 2B Instruct· Hugging Face Open LLM Leaderboard data15.3%
147Qwen1.5-14B-Chat· Hugging Face Open LLM Leaderboard data15.3%
148Ministral 8B· Epoch AI Benchmarking Hub data14.9%
149Claude 3 Haiku· Epoch AI Benchmarking Hub data14.9%
150Ministral 3B· Epoch AI Benchmarking Hub data14.4%

Showing 101–150 of 238 results.

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