Model positioning
OpenAI o3-mini is a lightweight reasoning model from OpenAI, optimized for technical tasks such as science, mathematics, and programming, with a focus on balancing reasoning capability, cost, and response speed. (openai.com)
Capabilities and uses
- Supports text input and output and can be used for tasks such as solving mathematical problems, scientific analysis, code generation, and debugging.
- Supports function calling, structured outputs, and streaming; low, medium, and high reasoning effort levels allow a trade-off between processing time and reasoning capability.
- The official model documentation lists a context window of 200,000 tokens and a maximum output of 100,000 tokens. (developers.openai.com)
Limitations and status
- Does not directly support image, audio, or video input; does not support fine-tuning. The knowledge cutoff date is October 1, 2023. Model outputs may still be incorrect, and important conclusions should be independently verified. (developers.openai.com)
- As of November 20, 2026, the official model documentation marks
o3-miniando3-mini-2025-01-31as shut down; the deprecation schedule lists October 23, 2026 as the shutdown date. (developers.openai.com)
Release date
January 31, 2025, the official initial release and API availability date. OpenAI published the model announcement that day, and the API changelog listed o3-mini and the version snapshot o3-mini-2025-01-31 on the same day, recording the model's launch. This date is not an end-of-month placeholder. (openai.com)
API pricing
The standard text token prices listed on the official model page are in U.S. dollars (USD): $1.10 per 1 million input tokens, $0.55 per 1 million cached input tokens, and $4.40 per 1 million output tokens. Cached input pricing applies only to inputs eligible for caching; these are the prices listed on the page and do not indicate that the model is currently available for use. Date checked: November 20, 2026. (developers.openai.com)