OpenAI claims a Millennium Prize solution, mathematicians push back

IEEE Spectrum reported on October 5, 2026, that OpenAI has claimed to solve one of the Millennium Prize problems, a claim met with both celebration and controversy. The same report goes back to the annual Heidelberg Laureate Forum in Germany, which opened on September 13 and brought together leading figures and up-and-coming talent in mathematics and computer science. There, mathematicians including Michael Harris, Peter Scholze, Geordie Williamson and Jacob Tsimerman took part in a debate specifically about how AI is affecting mathematical research. The coverage does not name which Millennium Prize problem is at issue, nor does it describe any proof.

A day later, The Verge widened the picture. Over the past year, it reported, OpenAI, Anthropic and other labs have announced breakthroughs on numerous long-standing mathematical problems, in some cases pushing well beyond what researchers expected current systems to be capable of — including resolving one of the famous Millennium Prize problems. The results have not brought unanimous applause. Instead they have triggered a backlash within the mathematics community, and OpenAI is now trying to repair relations by consulting an independent advisory group. The Verge characterized the labs' style as classic Silicon Valley: moving fast and breaking things as they barrel through the discipline.

Neither report states when or how OpenAI announced the claim, which problem was solved, or how any result has been verified. What is established at this point is the claim itself, the pushback from mathematicians, and OpenAI's effort to set up an independent advisory group.