OpenAI today officially launched their latest reasoning model, the o3-mini. For your information, this model was announced in December 2024, and now it is being offered to the public.
OpenAI says that the new o3-mini model offers reasoning support similar to the previous o1 model, but now it operates faster in providing feedback to users. This model is also marketed as more powerful and at the same time lower cost.
Under this reasoning model, it checks several times before providing feedback to users, thus avoiding various errors that occur in the final answer. It is also more suitable for problems involving programming, mathematics and science.
Interestingly this time, OpenAI is offering the o3-mini model to everyone, including free ChatGPT users. You just need to select the o3-mini model in the model menu, and then use it. Free version users are limited to generating answers multiple times, while paid users get a higher usage limit.