OpenAI Claims Millennium Prize Breakthrough Amid Plagiarism Allegations

OpenAI has claimed to solve the Navier-Stokes existence and smoothness problem, one of the most elusive challenges in mathematics. The announcement has triggered a sharp dispute with academic researchers who allege the company used its massive computational power and private user data to hijack their work.

According to OpenAI, the solution was achieved in just 88 hours using an unreleased AI model and approximately 10,000 agents. The company estimates the compute cost for the breakthrough at roughly $22.5 million. The Navier-Stokes problem is one of seven Millennium Prize Problems established by the Clay Mathematics Institute in 2000, which carries a $1 million reward for a verified solution.

A Clash of Discovery

The controversy centers on NYU Professor Tristan Buckmaster and Anthropic researcher Levent Alpöge, who had been pursuing their own solution. The researchers used various large language models, including OpenAI’s Codex, to develop preliminary proofs. The dispute erupted when OpenAI published a full proof shortly after Buckmaster made a preliminary announcement regarding his progress.

Buckmaster has questioned whether OpenAI’s models were trained on, or had access to, the private Codex sessions where he and Alpöge had stored their project drafts. "I asked whether the model had been trained on, or had access to, our sessions in Codex," Buckmaster stated, noting that he was told the model did not look up user data.

The Corporate Defense

OpenAI CEO Sam Altman denied allegations of plagiarism in a statement on X. Altman claimed that OpenAI had initially believed the other team had also solved the problem and attempted to coordinate a joint release. "When we learned that they had Euler but not Navier-Stokes, we offered to let them go first," Altman wrote.

Industry Implications

This incident highlights the growing ethical tension at the intersection of AI-driven scientific discovery and academic integrity. It raises critical questions about the definition of "independent discovery" when researchers rely on proprietary tools provided by the very companies competing with them. The case underscores the risk that AI developers could potentially leverage user data from their own platforms to outpace the scientists using those tools, creating a conflict of interest in high-stakes research.

What Remains Unclear

While OpenAI has announced its result, the mathematical community must still verify the proof to determine if it meets the rigorous standards of the Clay Mathematics Institute. It remains to be seen if the dispute over data privacy will lead to broader demands for transparency regarding how AI labs handle the private research sessions of their users.