OpenAI’s Math Win Sparks Theft Accusation From Scholar

OpenAI recently claimed a major milestone: solving one of the six unsolved Millennium Prize Problems set by the Clay Mathematics Institute, a set of math puzzles each carrying a million-dollar reward. The specific question, nearly a century old, concerns whether the equations describing fluid flow can ever break down under extreme conditions. According to OpenAI, the breakthrough came together in just a matter of days, powered by roughly 10,000AI agents working over a single weekend.

A Mathematician Pushes Back

Not everyone is celebrating. Mathematician Tristan Buckmaster has publicly disputed OpenAI’s account, telling Scientific American that he and collaborator Levent Alpöge, a researcher at rival lab Anthropic, had already developed substantial groundwork on the same problem before OpenAI’s announcement. Buckmaster says that once news of their unpublished approach circulated, OpenAI’s team used it to complete the proof before he and Alpöge could publish their own findings.

Credit Dispute Turns Contentious

The disagreement reportedly escalated from there. Buckmaster says OpenAI offered him full credit for the discovery, but only on the condition that Alpöge’s contribution be omitted entirely, since Alpöge works for a competing AI company. When Buckmaster resisted that condition, he says an OpenAI employee pressed him with a pointed question about jeopardizing his own career.

The episode raises fresh questions about how credit should be assigned when AI systems build on unpublished academic work, and about the competitive tensions now surfacing between rival AI labs racing to claim scientific firsts. OpenAI has not issued a detailed public response addressing Buckmaster’s specific allegations.