The company says its internal AI has solved the famous Navier-Stokes Millennium Prize problem
OpenAI has come under fire after claiming its AI solved one of mathematics’ most famous problems. Researchers are raising questions over how the company arrived at the breakthrough, while others are accusing it of exaggerating the achievement to boost valuation ahead of a potential IPO.
The company announced on Tuesday that an internal AI system had solved the Navier-Stokes problem, one of mathematics’ famous Millennium Prize questions. It concerns equations describing how liquids and gases move and asks whether the math can suddenly break down under certain conditions. A correct solution carries a $1 million prize from the Clay Mathematics Institute.
OpenAI said around 10,000 AI agents worked on the problem for roughly 88 hours, exchanging millions of messages and generating around 130 billion output tokens before producing a computer-verified proof.
The company says the result satisfies the official wording of the prize problem, although it does not intend to claim the money. It also presented the feat as evidence of how rapidly its internal AI systems are advancing.
The announcement has drawn mixed reactions, with some calling it an “exciting day for mathematics” while others have raised suspicions of plagiarism and questioned the significance of the achievement.
NYU mathematician Tristan Buckmaster has accused OpenAI of throwing massive resources at the problem only after learning that he and Anthropic researcher Levent Alpöge were making progress on a related approach.
Mathematicians have also cautioned against declaring the problem solved before the proof is independently examined. Diego Cordoba, whose work helped pioneer the approach used by the AI, said a valid solution would be a “big surprise.”
Others have been far more critical. Prominent tech commentator Mehdi, who has a large following on X, opened a lengthy attack on the announcement by telling OpenAI CEO Sam Altman to “go f**k yourself,” before accusing the company of trying to “privatize the very mechanism of human knowledge.”
Mehdi argued that OpenAI had tackled an easier version of the problem and used brute-force computing rather than genuine machine reasoning or scientific intuition.
He suggested that the supposed breakthrough was achieved by thousands of AI agents simply searching mathematical ground already explored by human researchers in order to manufacture an artificial “Sputnik moment” to boost excitement around its technology and inflate its valuation ahead of a stock-market listing.
OpenAI has acknowledged that the project began after hearing rumors that outside mathematicians had made progress on related problems but has denied accessing their unpublished work. At the same time, it said it could not rule out that anonymized data from their use of OpenAI products had helped improve its models.
