o1-mini
OpenAI · 2024-09-12
o1-mini is OpenAI's smaller, faster, cost-efficient reasoning model, released on September 12, 2024 and optimized for STEM reasoning, especially mathematics and coding. It was trained with the same high-compute reinforcement-learning pipeline as o1, was 80% cheaper than o1-preview at launch, and offers text-only input and output with a 128,000-token context window.