
As anticipated, OpenAI has officially rolled out its new o3-mini reasoning model, which will be accessible to all users starting this Friday. The level of access, however, will vary based on the user’s subscription tier.
OpenAI initially hinted at the o3 model family during the grand finale of its 12 Days of OpenAI livestream event last December, shortly after introducing the o1 reasoning model family. In mid-January, CEO Sam Altman revealed more details about the o3-mini model, ultimately confirming that it would be made available on the ChatGPT platform for all users.
Beginning Friday, users across all subscription levels can use the o3-mini model in ChatGPT. Free-tier users can activate it by clicking the “Reason” button in the prompt window, whereas paid subscribers can easily select it from a dropdown menu dedicated to model options.
Similar to other new features from OpenAI, users on the free tier will experience some limitations with the model. Plus and Teams subscribers will be allotted 150 queries per day, while Pro users, who pay $200 a month, will enjoy unlimited access. Additionally, paid users will have the privilege of utilizing o3-mini-high, which offers greater computational resources, albeit with slower response times.
Enterprise and education subscribers will gain access to this model next week, and it will also be available via the company’s API starting Friday.
The o3-mini model is distinct from conventional large language models (LLMs), as it internally verifies the accuracy and reproducibility of its outputs before presenting them to the user. This leads to longer inference times but results in more reliable answers. The o3-mini is particularly optimized for STEM-related tasks such as programming, mathematics, and scientific inquiries, providing comparable performance to the o1 model family while being more cost-effective and efficient.
“While the o1 model serves as our more general-purpose reasoning option, o3-mini offers a specialized alternative that excels in technical fields where precision and speed are vital,” OpenAI stated in its announcement. “The launch of o3-mini is another stride in OpenAI’s commitment to advancing affordable intelligence.”