Best AI Models for Math

Ranked from provider-published pricing · Prices checked November 17, 2026

The MATH benchmark uses high-school competition problems that require multi-step symbolic reasoning rather than recall. It separates models that can carry a chain of reasoning from those that pattern-match an answer.

The ranking

top 20 of 103

Ranked byMATH Benchmark Score. Prices checked November 17, 2026.

Gemini 2.0 Flash from Google leads this ranking at 90.9%. The median across the 103 qualifying models is 68%.

#ModelProviderMATH Benchmark ScoreGPQA DiamondInput /1MOutput /1MBlended 70/30
1Gemini 2.0 FlashGoogle90.9%65.2$0.1000$0.4000$0.1900
2Gemini 1.5 ProGoogle86%58.1$1.25$5.00$2.38
3Qwen2.5-MaxQwen85%60.1$1.60$6.40$3.04
4Qwen2.5-72B (Groq)Groq83.1%49$0.7900$0.7900$0.7900
5Qwen2.5-72B (Together)Together AI83.1%49$1.20$1.20$1.20
6Qwen2.5-72B (fw)Fireworks AI83.1%49$0.9000$0.9000$0.9000
7Qwen2.5-72B (DI)DeepInfra83.1%49$0.3500$0.4000$0.3650
8Qwen2.5-72B (Hyp)Hyperbolic83.1%49$0.4000$0.4000$0.4000
9Qwen2.5-72B-InstructQwen83.1%49$0.3600$0.4000$0.3720
10Qwen2.5-32B-InstructQwen83.1%49.5$0.7000$0.7000$0.7000
11Command ACohere80%50.8$2.50$10.00$4.75
12Phi-4Microsoft80%56.1$0.0700$0.1400$0.0910
13Qwen2.5-14B-InstructQwen80%—$0.3500$0.3500$0.3500
14Claude Sonnet 3.5Anthropic78.3%65$3.00$15.00$6.60
15Gemini 1.5 FlashGoogle78%50$0.0750$0.3000$0.1425
16MiniMax-Text-01MiniMax77.4%—$0.7000$2.80$1.33
17Llama 3.3 70BMeta77%50.5$0.2300$0.4000$0.2810
18Llama 3.3 70B (Groq)Groq77%50.5$0.5900$0.7900$0.6500
19Llama 3.3 70B (Together)Together AI77%50.5$0.8800$0.8800$0.8800
20Llama 3.3 70B (DI)DeepInfra77%50.5$0.2300$0.4000$0.2810

16 of the 20 scores shown link to the page they were read on (underlined; hover for publisher, variant and date read). The rest have no recorded source yet and are being worked through.

How this list is built

  • Ranked by MATH score, highest first. Scores are as published by the model’s provider or a public leaderboard. Where a hosted variant has no score of its own, it inherits the base model’s.
  • Includes models marked available or preview; retired and deprecated SKUs are excluded.
  • Prices are the rates each provider publishes, not estimates. How the blended rate is calculated.

182 otherwise-eligible models were left out because we hold no published value for the figure this list ranks by. They are excluded rather than assumed.

Frequently asked questions

Are AI models reliable at maths?

The strongest models score highly on competition-style problems, but a benchmark score is an average, not a guarantee, a model that solves 90% of MATH problems still fails one in ten. For anything where correctness matters, have the model show its working, or verify the result with a calculator or symbolic tool rather than trusting the answer directly.

Related rankings

Price it for your workload

Rankings are recomputed on every deploy from the catalogue. Prices checked November 17, 2026.