What exactly did OpenAI's model prove?

OpenAI's model generated a proof showing that a solution to the Navier-Stokes equation, a Millennium Prize Problem, can develop a finite-time singularity, meaning the fluid's velocity becomes infinite in finite time. The company says this establishes the 'C' and 'D' claims of the Clay Institute's problem.

Who is Tristan Buckmaster and what is he accusing OpenAI of?

Tristan Buckmaster is a mathematician at New York University. He accuses OpenAI of starting its work only after his own related research was communicated to them, and he questions the similarity in methods, suggesting their work may have been influenced. Sebastian Bubek of OpenAI disputes this as false.

Will OpenAI claim the $1 million prize money for this proof?

No, OpenAI states it does not intend to claim the $1 million prize money from the Clay Mathematics Institute.

Large Language ModelsITmedia AI＋1h agoSep 9, 2026, 13:00 JST2 min read

OpenAI claims AI solved Navier-Stokes; mathematician cries foul

Send on LINE

3 Key Points

What happened
On September 8, OpenAI announced its unreleased model produced a proof establishing singularities for the Navier-Stokes equation, a Millennium Prize Problem. The company said the proof satisfies the 'C' and 'D' claims of the Clay Institute's official problem.
Why it matters
Tristan Buckmaster, NYU mathematician, publicly accused OpenAI of starting its effort only after his own related results were communicated, citing similarities in method. Sebastian Bubek called this 'false and inflammatory', sparking a dispute over both the outcome and the process.
What to watch
The proof's validity hinges on independent verification by the mathematics community, which has yet to occur. OpenAI has stated it will not claim the $1 million prize.

WHO IT HITSMathematicians in fluid dynamics and related fields will need to rigorously verify the proof and the Lean formalization, determining if OpenAI's claim meets the Clay Institute's official criteria. The dispute also affects AI researchers and ethics reviewers who will assess the transparency and attribution of research processes.

Ask the AI about this article →

Summaries like this, in your inbox every morning.

Get news like this every morning — free

Context & Analysis

OpenAI's September 8th announcement represents a potential watershed for AI in mathematics: an unreleased internal model reportedly generated a proof that it claims addresses the Navier–Stokes equation, a Millennium Prize Problem. The company's approach, running about 10,000 agents for about 88 hours to reach a conclusion and investing millions of dollars in compute, signals an ambition to push beyond mere assistance into autonomous discovery.

The dispute, however, centers on process and originality. NYU mathematician Tristan Buckmaster, who, with Levent Alpoge, had independently published related results on finite-time blow-up, alleges that OpenAI's work began only after he communicated his own findings during a phone call on September 6th. Buckmaster points to his use of OpenAI's Codex and Anthropic's Claude as a potential vector for data leakage, a claim OpenAI denies while not entirely ruling out the influence of anonymized user data. The clash highlights the blurred lines between human and machine contribution and the critical importance of data provenance in an era of AI-assisted research.

Validation remains the ultimate test: the mathematics community must independently review both the proof and its formalization in the proof assistant Lean to confirm it aligns with Clay's official problem setting. If verified, it would mark a significant first, far surpassing any previous AI-generated achievement in this domain. The unresolved dispute, meanwhile, echoes as a warning on research ethics, possibly influencing how AI research is conducted and credited in the future.

FAQ

What exactly did OpenAI's model prove?: OpenAI's model generated a proof showing that a solution to the Navier-Stokes equation, a Millennium Prize Problem, can develop a finite-time singularity, meaning the fluid's velocity becomes infinite in finite time. The company says this establishes the 'C' and 'D' claims of the Clay Institute's problem.
Who is Tristan Buckmaster and what is he accusing OpenAI of?: Tristan Buckmaster is a mathematician at New York University. He accuses OpenAI of starting its work only after his own related research was communicated to them, and he questions the similarity in methods, suggesting their work may have been influenced. Sebastian Bubek of OpenAI disputes this as false.
Will OpenAI claim the $1 million prize money for this proof?: No, OpenAI states it does not intend to claim the $1 million prize money from the Clay Mathematics Institute.

ITmedia AI＋Read Original Article

Send on LINE