OpenAI releases AI-generated math results with Lean formalizations
OpenAI is publicly releasing a broad range of new mathematical results produced by an internal unreleased frontier model, after consulting the independent Advisory Group on Mathematics and Artificial Intelligence at the Institute for Advanced Study. The work is posted to a GitHub repo with paper-revision and citation protocols; many proofs are formalized in Lean, with more to come, plus 10 reasoning summaries, problem-attempt stats, and compute estimates averaging roughly three hours of ChatGPT Pro thinking per result. OpenAI plans to fund workshops and conferences on the findings and eventually release the model itself in a responsible manner. This is a coordinated attempt to share original AI-generated research with the mathematics community rather than keep it proprietary.