Best AI models for math
AI models ranked on math benchmarks: MATH, AIME, and advanced arithmetic reasoning.
Math-capable models solve competition-style problems and symbolic derivations reliably.
- 1
Phi-4
86.5FrontierMicrosoft- Context:
- 16K
MathFrontierOpen source - 2
DeepSeek V3
84.0FrontierDeepSeek- Context:
- 128K
- Input:
- $0.27/1M
- Output:
- $1.1/1M
MathAgenticFrontierOpen sourceCode - 3
o3
83.5FrontierOpenAI- Context:
- 200K
- Input:
- $2/1M
- Output:
- $8/1M
VisionMathAgenticLong contextFrontierReasoning - 4
Gemini 2 Pro
82.8FrontierGoogle- Context:
- 2M
- Input:
- $1.25/1M
- Output:
- $5/1M
VisionMathAgenticLong contextFrontierCode - 5
Grok 3
80.1FrontierxAI- Context:
- 131K
- Input:
- $3/1M
- Output:
- $15/1M
VisionMathAgenticFrontierReasoning - 6
GPT-5
77.9StrongOpenAI- Context:
- 272K
- Input:
- $1.25/1M
- Output:
- $10/1M
VisionMathAgenticLong contextReasoningCode - 7
GPT-4o mini
77.1StrongOpenAI- Context:
- 128K
- Input:
- $0.15/1M
- Output:
- $0.6/1M
VisionMathAgenticBudget - 8
DeepSeek R1
76.9StrongDeepSeek- Context:
- 128K
- Input:
- $0.55/1M
- Output:
- $2.19/1M
MathOpen sourceReasoning - 9
o1
76.3StrongOpenAI- Context:
- 200K
- Input:
- $15/1M
- Output:
- $60/1M
VisionMathAgenticLong contextReasoning - 10
o3-mini
74.5StrongOpenAI- Context:
- 200K
- Input:
- $1.1/1M
- Output:
- $4.4/1M
MathAgenticLong contextCode