OpenAI announced on Tuesday that it has published a solution to the Navier-Stokes existence and smoothness problem, one of the seven legendary Millennium Prize Problems in mathematics. The claim, which represents the first AI-generated solution to one of these historic challenges, has immediately sparked a fierce academic and corporate dispute. The Navier-Stokes equations, which describe how fluids move, are fundamental to mathematical physics but have remained poorly understood in theoretical terms for generations.
NYU mathematics professor Tristan Buckmaster, who was simultaneously preparing to announce his own preliminary findings on the same problem, has accused OpenAI of using leaked information and its massive computing power to front-run his research. The clash exposes the intense rivalry between commercial AI labs and traditional academic researchers as artificial intelligence increasingly encroaches on high-level theoretical mathematics.
The Million-Dollar Millennium Challenge
Established in 2000 by the Clay Mathematics Institute, the Millennium Prize Problems are a set of seven of the most profound and difficult unsolved questions in mathematics, with each carrying a $1 million bounty for a successful proof. Prior to Tuesday's announcement, only one of these problems—the Poincaré Conjecture, solved by Grigori Perelman in 2003—had ever been resolved. OpenAI's claimed breakthrough involved spending millions of dollars in computing resources to generate a formal proof.
However, the achievement was instantly overshadowed by Buckmaster's public statement, which detailed an unusual sequence of events leading up to OpenAI's publication. Buckmaster had been working on the problem alongside Levent Alpöge, a mathematician employed by rival AI lab Anthropic who was conducting the research in a personal capacity. The duo relied on a mix of AI models, primarily OpenAI's Codex, to assist in assembling their proofs.
Accusations of Academic Front-Running
According to Buckmaster, he and Alpöge were finalizing three proofs when they discovered that "information about our progress had been passed to OpenAI." Upon contacting the AI giant, they were initially told that OpenAI had already achieved a full proof of the central problem. However, Buckmaster claims that when they pressed for details regarding when OpenAI's research began and the extent of human intervention, the company's responses grew evasive.
"It emerged that an entire team had been working on the problem," Buckmaster wrote in his statement, "and that an insane amount of compute had been used…. Eventually, it was agreed that [the first prompt] had been sent in the past few days, after information about our work had reached OpenAI." If these allegations are accurate, they suggest that OpenAI recognized the validity of Buckmaster and Alpöge's specific mathematical approach and deployed its vast computational infrastructure to race ahead and claim the formal proof first.
A Rare Mathematical Pathway and Codex Concerns
The dispute is further complicated by the highly specific mathematical route chosen by the researchers. While the Navier-Stokes problem is widely studied, Buckmaster and Alpöge utilized a rare tactic involving a "smooth force," corresponding to options c and d in mathematician Charles Fefferman's official statement of the problem. Buckmaster noted that "almost nobody else I know of was working on it," adding that "it is not the direction one arrives at in a few days by giving a model the problem statement."
Buckmaster raised concerns regarding the tools used in his own research. Because the duo relied primarily on OpenAI's Codex model to assemble their project, and because OpenAI reserves the right to train its models on Codex user interactions unless users explicitly opt out, Buckmaster suggested it is plausible that OpenAI's systems may have ingested his work and regurgitated the methodology when prompted by OpenAI's internal team. OpenAI has not responded to requests for comment regarding this possibility.
Threats and Denials in the Research Community
The personal interactions between the researchers and OpenAI executives have also come under scrutiny. Buckmaster alleged that Sebastian Bubeck, the head of mathematical research at OpenAI, asked him to remove Alpöge's name from the credits as part of a proposed compromise, potentially due to Alpöge's employment at rival Anthropic. When Buckmaster insisted on making the dispute public, he claims Bubeck asked him, "Why would you ruin your career?" and later warned, "If you don’t want me to be nice, then I don’t have to be nice."
Bubeck has strongly rejected these allegations, describing Buckmaster's claims as "false and inflammatory" in a public post. "To clarify, I came into the discussion following academic norms, and I’m disappointed that it has come to this," Bubeck wrote, adding that "academic standards are of the highest importance to me" and promising to release a more complete statement in the near future.
Peer Review and the Two-Year Verification Window
The mathematical community must now begin the rigorous process of peer-reviewing both OpenAI's claimed proof and the preliminary findings published by Buckmaster and Alpöge. Because the Clay Mathematics Institute requires any solution to a Millennium Prize Problem to withstand two years of academic scrutiny and publication in a reputable journal before the $1 million prize can be awarded, a final determination on the validity of either proof remains years away.
Buckmaster has called for full transparency, stating that he has not yet seen OpenAI's proof and wants all research data made public to ensure an objective evaluation. "I have not seen OpenAI’s proof," Buckmaster wrote. "I do not know what their model did, or how. I do not know whether our data was used. I am not accusing anyone of anything. I am stating what I was told, when, and what was proposed to me. I am stating it because the alternative is to let a sequence of announcements say something I know to be false."
