In a September 21, 2026 post, OpenAI said it began training a new internal model on August 28 that, according to the company, resolved the Navier–Stokes Millennium Prize problem and has since resolved more than 100 long-standing open problems across most areas of mathematics. OpenAI said the pace of the model's progress surprised its own mathematicians and prompted internal discussion about how to inform the field. These claims come from OpenAI's own announcement and have not been independently verified in the post. OpenAI referenced a recent open letter, 'A Severe Misalignment of AI in Mathematics,' in which mathematicians raised concerns about using open problems as benchmarks for new AI systems. In response, OpenAI said it is working with mathematicians who have formed an independent advisory group, hosted at the Institute for Advanced Study, to serve as a bridge to the mathematical community. According to OpenAI, the group will advise on reviewing and communicating emerging results, assessing their significance, coordinating dissemination, and how tools can support research and learning. OpenAI stated the group operates independently, its members are unpaid, it can make advice public and change its membership, and it will not advise on the pace of internal progress. Initial named members include Timothy Gowers (Collège de France, Cambridge) and Martin Hairer (EPFL, Imperial College London).
- OpenAI says an internal model, in training since August 28, resolved the Navier–Stokes Millennium Prize problem and 100+ open math problems (per OpenAI, not independently verified).
- An independent mathematics advisory group, hosted at the Institute for Advanced Study, will advise on reviewing and communicating results but not on pacing OpenAI's internal work.
- Initial members include Timothy Gowers and Martin Hairer; members are unpaid and the group can make its advice public.
What it means for you
OpenAI claims one of its unreleased models has solved major unsolved math problems, including a famous prize problem, and is setting up a group of mathematicians to help vet and communicate those results. For now these are the company's own claims about an internal model, not something you can use or independently confirm. Nothing here changes any tool you have access to today.
Who should care
Working mathematicians, academic researchers, and people following the debate over how AI research results get verified and announced.
Skip this if
You run a business or use AI for everyday work — there is nothing to act on, and none of this is a product you can touch.
Sources: OpenAI — read the original