Back in December 2023, OpenAI first announced the new o3 series of reasoning models during its '12 Days of OpenAI' event. The o3 series includes o3 and o3 mini models. Today, OpenAI announced the availability of o3-mini models in ChatGPT and the API for developers.
As announced before, the o3-mini model will provide an option for users to select between three reasoning levels: High, Medium, and Low. The Low level will be the fastest but less accurate, while the High level will be the slowest but more accurate.
Until now, OpenAI's o1 series reasoning models were available only for ChatGPT paid tier users. With o3-mini, OpenAI is bringing the reasoning capabilities to free ChatGPT users as well. OpenAI claims that o3-mini with medium can match the performance of the OpenAI o1 model in math, coding, and science. The main highlight here is that the o3-mini is able to deliver it at a significantly lower cost and also with less time.
ChatGPT Plus, Team, and Pro users can access OpenAI o3-mini starting today, and Enterprise plan users can access it next week. OpenAI will be limiting these users to 150 messages per day with o3-mini. ChatGPT free users can use o3-mini using the new ‘Reason’ option in the message composer or by regenerating a response. Also, for the first time, OpenAI has integrated web search content with its reasoning models. So, o3-mini can also deliver up-to-date answers with links to relevant web sources.
For developers, o3-mini is now available with full support for function calls, Structured Outputs, streaming, and developer messages. Also, developers can access it via the Chat Completions API, Assistants API, and Batch API. o3-mini is also significantly cheaper when compared to the o1 series models. OpenAI claims that o3-mini will be 63% cheaper than OpenAI o1-mini and 93% cheaper than o1.
The o3 mini model's availability marks a significant step in democratizing access to advanced reasoning capabilities to both general consumers and developers.
1 Comment - Add comment