OpenAI surprised the world yesterday afternoon by announcing not "Strawberry" as rumored, nor GPT-5, but a new family of "reasoning" large language models (LLMs) called o1 that aims to offer high ...