OpenAI has introduced a new AI model called “o1,” which the company claims represents a significant advancement in reasoning capabilities for AI systems. This model, designed to solve complex problems in fields like science, mathematics, and coding, mimics human-like reasoning by taking a more thoughtful, step-by-step approach to problem-solving. For instance, in testing, the o1 model performed on par with PhD students in solving difficult tasks across disciplines such as physics and biology.
The o1 model is part of OpenAI’s broader “Strawberry” project, which aims to push the boundaries of AI reasoning and decision-making. Compared to previous versions like GPT-4, o1 demonstrates enhanced accuracy and fewer hallucinations (incorrect responses), making it more reliable for complex tasks. OpenAI CEO Sam Altman acknowledges that while the model is still imperfect, it marks the beginning of a new era in AI development—one where models are capable of more generalized, sophisticated reasoning.
The o1 model is available for ChatGPT Plus and API users, with a lighter and more affordable version, o1-mini, tailored for developers, offering a balance between performance and cost-effectiveness
