History
o3-mini brings a smaller reasoning model
OpenAI introduced o3-mini in ChatGPT and the API as a cheaper model for reasoning tasks.

On 31 January 2025, OpenAI released o3-mini in ChatGPT and the API. The model focused on reasoning tasks, including mathematics and programming. Developers could choose how much reasoning effort the model spent on an answer. The introduction describes these capabilities.

For an application handling many questions, answer quality, waiting time and cost each deserve their own consideration. The question was not just which model was available, but also how much effort an answer should receive. A smaller model therefore deserved its own place in a comparison.

For a small or medium-sized business, one possible trial is checking quantities in a simple stock calculation. Supply practice data and ask which lines contain calculation errors. Prepare an answer sheet yourself beforehand. Use it to count both missed errors and false warnings. Then compare different settings using exactly the same assignments. Record how long each answer takes and how much review remains necessary. Use the findings to choose an appropriate approach. A more expensive or slower setting need not be the default; let the task determine the required quality.