o1-mini

Available frontier

Fast reasoning for coding and STEM

Released September 12, 2024

API Documentation View Provider

Context Window

128K

TTFT

N/A

Speed

N/A

Max Output

65,536

Training Cutoff

Oct 2023

Last tested: 2026-01-15

About

o1-mini is a faster, more affordable reasoning model optimized for coding, math, and science tasks. It delivers 80% of o1's reasoning capability at 80% lower cost, making it ideal for developers who need strong reasoning without the full o1 price tag.

Capabilities

function-calling streaming extended-thinking reasoning-tokens

Pricing

Input: $3.00/M
Output: $12.00/M

Live pricing Static pricing

Batch

$1.50/M

per 1M tokens

Details

Overview

o1-mini brings the reasoning revolution to cost-conscious developers. Released alongside the original o1-preview, it’s specifically optimized for STEM reasoning tasks where broad world knowledge is less critical than logical problem-solving ability.

Key Features

Efficient Reasoning: Focused reasoning capability optimized for technical tasks
Fast Thinking: Significantly faster response times compared to full o1
Cost Effective: 80% cheaper than o1 while retaining strong performance
STEM Optimized: Particularly strong at coding, math, and scientific reasoning
Large Output: Support for up to 65K output tokens

Use Cases

o1-mini is ideal for code generation and debugging, algorithmic problem solving, mathematical proofs, scientific analysis, and automated code review. It’s the go-to choice for developers building AI-powered coding assistants.

Performance

o1-mini matches or exceeds o1-preview on many coding and math benchmarks while being significantly faster and cheaper. It achieves strong results on competitive programming problems and complex debugging tasks.