OpenAI claims AI model solved major mathematics problem, but credit dispute deepens
OpenAI says an unreleased model solved the Navier–Stokes problem in 88 hours, while researchers dispute the company’s account and the result awaits formal scrutiny.
OpenAI says an unreleased artificial intelligence model has solved the Navier–Stokes existence and smoothness problem, one of seven Millennium Prize Problems established by the Clay Mathematics Institute.
The company said the model reached its result after about 88 hours, with 10,000 AI agents working on the problem and computing costs running into millions of dollars. OpenAI said the work showed the equations describing fluid motion were fatally flawed.
The claim, reported by France 24, has been accompanied by a dispute over the origins of the approach. New York University mathematician Tristan Buckmaster and Levent Alpoge, a researcher at rival Anthropic, allege OpenAI pursued their method after learning of their AI-assisted work.
OpenAI researchers deny accessing the pair’s work and say they did not use their prompts or proofs to direct the models. The company has nevertheless said it cannot rule out the possibility that de-identified data from their use of its products contributed to model improvement.
The result has not been independently validated. Under the Clay Mathematics Institute’s rules, any proposed solution must be published, undergo peer review and survive two years of scrutiny before the institute considers a prize claim.
Each Millennium Prize Problem carries a US$1 million reward. OpenAI says it will not seek the money if its finding is confirmed, while the broader claim that AI-generated mathematics demonstrates genuine understanding remains contested.
