Explore o3-mini: An initial look

O3-mini has been released by OpenAI on January 31, 2025. It is designed to provide cost-effective and efficient performance, particularly in STEM fields such as science, math, and coding. o3-mini is available through both ChatGPT and the API.

Depending on your access level and needs OpenAI’s o3-mini can be used in multiple ways as given below:

1. Using o3-mini in ChatGPT

• If you're using ChatGPT (Free, Plus, Team, or Pro):

  • Free Users: You can access o3-mini by selecting the “Reason” mode in the message composer.

  • Plus/Team/Pro Users: The model is automatically available, but Pro users can switch between different models, including o3-mini.

2. Using o3-mini via the OpenAI API

• You can call o3-mini in the OpenAI API by specifying it as the model in your request:

json

{

"model": "o3-mini",

"messages": [{"role": "user", "content": "What is quantum computing?"}]

}

• The API supports function calling, structured outputs, and reasoning effort levels (low, medium, high).

3. Using o3-mini in Third-Party Apps

If you use platforms that integrate OpenAI models (e.g., Notion AI, Zapier, or chatbots), check if they have added support for o3-mini.

4. Experimenting with Reasoning Effort

• Low effort → Faster, but may be less precise.

• Medium effort → Balanced speed and accuracy.

• High effort → More detailed reasoning but slower responses.