Benchmark Scores
89.5/100MMLU86.9
HumanEval90
Math90.2
Reasoning90.5
Coding89.8
About OpenAI o3-mini
OpenAI o3-mini is a cost-efficient reasoning model optimized for STEM tasks. It offers adjustable reasoning effort levels (low, medium, high) and provides strong performance on math and coding at a fraction of o1's cost.
Strengths
- +Cost-effective reasoning
- +Strong math and coding at low cost
- +Adjustable reasoning effort
- +Fast for a reasoning model
Weaknesses
- -Narrower than full o1 on some tasks
- -Text-only input
- -Less creative than GPT-4o
Pricing
Per 1M tokens
Input$1.10
Output$4.40
Quick Facts
- Context Window
- 200K
- Parameters
- Undisclosed
- Release Date
- 2025-01-31
- Category
- Frontier
Compare Models
See how OpenAI o3-mini stacks up against other models