OpenAI recently announced the launch of o3-mini for OpenAI API and ChatGPT users.
The company initially unveiled o3-mini in 2024, with the more cost-efficient model acting as a direct competitor versus DeepSeek. The developers said the AI model ‘advances the boundaries’ and is designed for STEM functions- OpenAI said that it’s good in coding, math, and science. The model is also search-capable and can show web source links and find up-to-date answers.

OpenAI said o3-mini will be for technical domains while keeping o1 as the ‘standard general knowledge’ platform. O3-mini users can pair it with ChatGPT for medium reasoning effort and a higher option in the model picking selection. It’s worth noting that o3 is the first reasoning model that free users can access. Pro, Team, and ChatGPT Plus subscribers can access the o3-mini model right away and replace the o1-mini as the default. The limit has also been raised to 150 from 50.