OpenAI Launches o1 Reasoning Model, Touting Major Gains in Math and Science
Read the full article for context, quotes, and updates from the team.
OpenAI has unveiled o1, a new family of reasoning models built to tackle complex problems in mathematics, coding and science. The company says the model marks a major step forward in AI performance by using reinforcement learning to improve how it approaches difficult tasks, effectively “thinking” through problems step by step before generating an answer.
According to OpenAI, o1 delivered strong results on several challenging benchmarks. The model scored 83% on AIME 2024, a test of advanced math reasoning, and 74.4% on GPQA Diamond, a benchmark designed to measure graduate-level science knowledge and problem-solving ability. OpenAI said these results reflect the model’s ability to reason more carefully than previous systems on tasks that require multi-step analysis.
The o1 family is now available to ChatGPT Plus and Team users, with a wider rollout planned later. OpenAI positioned the release as an important milestone in the development of AI systems that can handle more demanding real-world work, particularly in technical fields where accuracy and reasoning are critical.
The launch adds to growing competition in the AI industry as companies race to build models that go beyond fluent text generation and toward more reliable problem-solving capabilities.
