o1

O1 is OpenAI's reasoning-focused model, trained with reinforcement learning to reason deeply before responding. It produces a long internal chain of thought, spending substantial computation on problem-solving rather than immediate generation. The model accepts text and image input and produces text output, with up to 100,000 output tokens. It supports function calling, structured outputs, streaming, and the Responses API for applications that need extended reasoning. O1 is particularly strong in physics, chemistry, biology, mathematics, and software engineering, where multi-step reasoning provides the greatest advantage. With a 200,000 token context window, it handles substantial problem specifications and long documents.