OpenAI is set to launch its new AI model, o3-mini, in the coming weeks, according to Chief Product Officer Kevin Weil. The o3-mini will feature advanced reasoning capabilities in science, code, and mathematics. Following the mini version, the full o3 model is expected to be released in February or March, contingent on development progress. Additionally, OpenAI is already training the next model, referred to as o4, which is anticipated to offer significant advancements in capabilities. OpenAI's iteration cycle for AI models is reportedly accelerating, with o3 now recognized as the 175th best software engineer globally. This rapid development comes as the AI landscape remains competitive, with recent releases from other companies, including Google's gemini-2.0-flash-thinking-exp-01-21.