September 12, 2024: OpenAIs o1 Model Excels in Reasoning and Math - OpenAIs new o1 LLM series, codenamed Strawberry, excels in reasoning, decoding scrambled texts, and acing difficult exams, outperforming earlier models like GPT-4o. The series features two models: o1-preview, noted for its high accuracy, and o1-mini, which offers cost-efficiency. Key advancements include the use of CoT (chain of thought) and reinforcement learning, contributing to superior performance in complex tasks and enhanced safety. The models are accessible through paid ChatGPT services and APIs, with plans for wider availability and higher usage limits in the future.