OpenAI launches o3-mini, its new and smarter reasoning model for paid & free ChatGPT users

OpenAI has introduced o3-mini, its new reasoning model designed to enhance speed and accuracy compared to the previous o1 model. This marks the first time that free ChatGPT users can access a reasoning model, albeit with rate limits akin to GPT-4o. Originally announced in December, o3-mini is optimized for tasks in math, coding, and science, delivering responses 24% faster than o1 while maintaining performance levels.

A key feature of o3-mini is its ability to explain the reasoning behind its answers, a step beyond simply providing results. Benchmarks from December showed o3 outperforming o1 in reasoning and coding tasks, and OpenAI confirms that o3-mini surpasses o1 with reduced cost and latency. Developers can integrate o3-mini via OpenAI’s Chat Completions API, Assistants API, and Batch API.

For those on paid plans, o3-mini-high is available, promising the "best coding option in ChatGPT," with improved intelligence but slightly longer response times. This model also features search capabilities, allowing it to deliver responses with linked web sources. OpenAI is also increasing message limits for ChatGPT Plus and Teams users to 150 messages per day, while Pro users, who pay $200/month, will have unlimited access to o3-mini. This development comes on the heels of the recent launch of DeepSeek-R1, a competitive Chinese reasoning model that offers powerful performance at a lower cost.