OpenAI launched o3-mini on Friday, a cost-efficient AI model optimized for reasoning in science, technology, engineering, and math (STEM).
The company stated that this is the only model that supports developer features such as function calling, structured outputs, and developer messages, making it ready for production. The o3-mini will also support streaming. "This flexibility allows o3-mini to 'think harder' when tackling complex challenges or prioritize speed when latency is a concern," the company added.
The new model outperforms previous versions in coding and mathematics, with testers preferring its responses 56% of the time. OpenAI has also expanded access by tripling message limits for paid users and allowing free users to try o3-mini via the "Reason" option in ChatGPT.