LLM Math Reasoning Benchmark Leaderboard

Name: LLM Math Reasoning Benchmark Leaderboard
Creator: DataLearner
License: https://creativecommons.org/licenses/by/4.0/

This page provides the most comprehensive LLM math reasoning benchmark leaderboard. We evaluate models including GPT, Claude, Qwen, and DeepSeek using authoritative math benchmarks such as AIME 2025, FrontierMath-Tier4, MATH-500, and GSM8K.

Updated on 2026-05-02 07:14:49

As of 2026-05, this page covers AIME2025, FrontierMath - Tier 4, MATH-500, GSM8K and related benchmarks for LLM Math Reasoning Benchmark Leaderboard, making it straightforward to compare within the same task family.

Click any model name to check context length, licensing, and pricing on its detail page. See Data Methodology for scoring details.

Benchmark

AIME2025 FrontierMath - Tier 4 MATH-500 GSM8K

More Benchmarks

Model Size:All 3B and below 7B 13B 34B

LLM Performance Results

Data source: DataLearnerAI

Rank	Model					License
	Qwen3-4B-Thinking-2507 阿里巴巴	81.30	—	—	—	Free commercial
	Hunyuan-7B Tencent ARC	75.30	—	93.70	—	Free commercial
	Qwen3-8B 阿里巴巴	67.30	—	97.40	—	Free commercial
4	Qwen3-4B-2507 阿里巴巴	47.40	—	—	—	Free commercial
5	Pangu Embedded 华为	—	—	92.40	95.98	Free commercial
6	Qwen2.5-7B 阿里巴巴	—	—	—	85.40	Free commercial
7	Llama3.1-8B-Instruct Facebook AI研究实验室	—	—	—	82.40	Free commercial
8	Gemma 2 - 9B Google Research	—	—	—	70.70	Free commercial
9	Llama3.1-8B Facebook AI研究实验室	—	—	—	55.30	Free commercial
10	Mistral-7B-Instruct-v0.3 MistralAI	—	—	—	36.20	Free commercial
11	DeepSeek-R1-Distill-Qwen-7B DeepSeek-AI	—	—	91.40	—	Free commercial