#21
o3-mini
o3-mini

OpenAI's o3-mini is a highly efficient reasoning model released in early 2025, specifically optimized for STEM tasks like coding, mathematics, and science. It delivers state-of-the-art performance on reasoning benchmarks at a fraction of the cost and latency of larger models. While it trades off some general creative writing ability, its instruction following and logical analysis capabilities are exceptional, making it a powerful tool for technical problem solving and agentic workflows.

Performance Metrics

Helpfulness
Instruction Following
Comprehension
Empathy
Creative Writing
83.60
Helpfulness
88
Empathy
75
Instruction Following
94
Creative Writing
72
Comprehension
89
Speed
Avg 92 tok/s

Model Specifications

Release Date
January 31, 2025
Lab
OpenAI
Type
Proprietary
Context Size
200K
Max Output Tokens
100K
Cost per 1M tokens
$1.10 / $4.40
Model Inputs
Text
Model Outputs
Text
Tool Calling
Enabled

Compare With Similar Models