OpenAI says an unreleased AI model has produced a solution approach to the 90-year-old Navier–Stokes existence and smoothness problem, one of the seven Millennium Prize Problems. The company describes the effort as taking 88 hours, using about 10,000 AI agents operating in parallel, exchanging millions of messages and generating a very large volume of output tokens. OpenAI says the work addresses two of the four required statements for a full solution.
Several outlets report that the claim is not yet accepted or verified by independent experts. OpenAI emphasizes that its goal is to demonstrate AI capabilities rather than to pursue prize money. Times of India also notes criticism from NYU professor Tristan Buckmaster, who alleges the approach resembled an unpublished route that he had been developing and shared via drafts that may have been incorporated into Codex. OpenAI denies using user data, while stating it cannot rule out that de-identified training materials may have influenced improvements to its models. Independent validation is still pending, and the Clay Mathematics Institute has not accepted the result.
