Chain Of Thought

OpenAI o1: Reasoning Models

OpenAI released o1, a model trained to 'think before it speaks' using chain-of-thought reasoning at inference time. It could solve complex math, coding, and science problems by spending more compute thinking through multi-step solutions — trading speed for accuracy on hard problems.