AI that can do general-purpose complex reasoning | OpenAI o1

A new large language model trained with reinforcement learning to perform complex reasoning. o1 thinks before it answers—it can produce a long internal chain of thought before responding to the user.