Best AI Models for Math

AI models ranked by mathematical reasoning using MATH-500, GSM8K, and AIME 2024 benchmark scores.

Last updated: 58m ago

#1 Model

Score: 88.2

Average Score

72.1

Across all ranked models

Models Ranked

With benchmark data

Weights:MATH-500 (40%)GSM8K (30%)AIME 2024 (30%)

Top Best for Math Models by Weighted Score

Top 15 models by weighted score

LMMarketCap.com

Benchmark Breakdown

Per-benchmark scores for top 10 models

MATH-500

GSM8K

AIME 2024

LMMarketCap.com

#	Model	Provider	Score	MATH-500	GSM8K	AIME 2024
1	o3OpenAI	OpenAI	88.2	99	--	96.7
2	o4 MiniOpenAI	OpenAI	86.1	97.3	--	93.4
3	Gemini 2.5 ProGoogle	Google	84.4	95.2	--	92
4	o3 MiniOpenAI	OpenAI	84	97.9	--	87.3
5	R1 0528DeepSeek	DeepSeek	82.7	97.8	--	84
6	o1OpenAI	OpenAI	81.7	96.4	--	83.3
7	R1DeepSeek	DeepSeek	80.8	97.3	--	79.8
8	Gemini 2.5 FlashGoogle	Google	78.1	85.8	--	88
9	DeepSeek V3DeepSeek	DeepSeek	76.9	90.2	96.7	39.2
10	Claude Opus 4.6Anthropic	Anthropic	76.6	90.5	--	78
11	GPT-5.4OpenAI	OpenAI	76.4	95.5	--	--
12	GPT-4oOpenAI	OpenAI	76.3	76.6	95.8	--
13	GPT-5.2OpenAI	OpenAI	75.2	94	--	--
14	GPT-5.1OpenAI	OpenAI	74.8	93.5	--	--
15	GPT-5OpenAI	OpenAI	74	92.5	--	--
16	GPT-4 TurboOpenAI	OpenAI	73.7	72.6	94.2	--
17	DeepSeek V3 0324DeepSeek	DeepSeek	73.6	92	--	--
18	Claude Opus 4.5Anthropic	Anthropic	73.1	88.1	--	72
19	GPT-4o-miniOpenAI	OpenAI	72.1	70.2	93.2	--
20	Llama 3.1 70B InstructMeta	Meta	71.7	68	95.1	--
21	Gemma 4 31BGoogle	Google	71.4	--	--	89.2
22	Claude Opus 4Anthropic	Anthropic	70.5	86	--	68
23	Gemini 3 Flash PreviewGoogle	Google	70.4	88	--	--
24	Claude Sonnet 4.6Anthropic	Anthropic	68.9	85.3	--	65
25	Gemma 2 27BGoogle	Google	68.2	--	85.3	--
26	Claude Sonnet 4.5Anthropic	Anthropic	65.8	83	--	60
27	Llama 4 MaverickMeta	Meta	64.8	81	--	--
28	Phi 4Microsoft	Microsoft	64.3	80.4	--	--
29	Claude Sonnet 4Anthropic	Anthropic	64.2	81.4	--	58
30	GPT-4.1OpenAI	OpenAI	62.8	78.5	--	--
31	Llama 3.3 70B InstructMeta	Meta	61.6	77	--	--
32	Mistral LargeMistral AI	Mistral AI	60.8	76	--	--
33	Claude Haiku 4.5Anthropic	Anthropic	58	72.5	--	--
34	Llama 4 ScoutMeta	Meta	40.2	50.3	--	--

How scores are calculated

Each model's score is a weighted average of its available benchmark results. When a model is missing some benchmarks, the weights are re-normalized across the benchmarks that are available. All scores are on a 0-100 scale. Data sourced from official model cards, published papers, and third-party evaluation platforms.

Other Specialty Leaderboards

Best for Coding Best for Reasoning Best for Writing Best for Instructions Best for Data Analysis Best for Roleplay Best for Multilingual

Frequently Asked Questions

Based on our benchmark analysis, o3 by OpenAI is currently the #1 ranked model for math, with a weighted score of 88.2/100.

Models are ranked using a weighted average of MATH-500, GSM8K, AIME 2024 benchmark scores. All scores are normalized to a 0-100 scale.

We currently rank 34 models that have relevant benchmark data for math tasks.