OpenAI says an unreleased experimental system produced a proposed solution to the Navier-Stokes existence and smoothness problem. It is one of the Millennium Prize Problems, and the prize is $1 million. The company claims around 10,000 AI agents worked together for about 88 hours. That is a striking claim, and it is already drawing scrutiny.
What OpenAI says happened
OpenAI describes a system more capable than GPT-6 Astra, its most advanced public model. The agents exchanged roughly 2.7 million messages and generated about 130 billion tokens while working on the problem. After that, Astra spent 17 more hours formalizing and checking the proof, according to OpenAI.
The proposed proof says a fluid vortex can become stretched and concentrated until its speed blows up in finite time. That would mean smooth flow can break down, even if total energy stays finite. The Navier-Stokes equations describe how fluids like water and air move. The open question asks whether a perfectly smooth flow always remains smooth, or whether it can develop unlimited speeds.
Real fluids cannot move infinitely fast under current physics. At extreme scales, the assumption that a fluid is one continuous substance stops matching reality. So the result may say more about where the equations fail than about any real-world flow. I think that distinction matters.
Why the math matters
Navier-Stokes equations are used in aircraft design, weather forecasting, and blood-flow research. A better understanding of where they break down could help scientists model extreme and turbulent flows. But the practical impact will take time. No one should expect overnight changes to engineering software.
The bigger implication may be the system that found the proof. If thousands of AI agents can attack a 90-year-old math problem in days, similar methods could be aimed at hard problems in materials science, energy, aerospace, and medicine. That is a big if, though. Verification is not the same as acceptance.
The fight over credit
There is controversy over how independently the result was reached. NYU mathematician Tristan Buckmaster and Anthropic researcher Levent Alpöge were separately working on a closely related fluid-dynamics problem and had used AI tools. Buckmaster questioned on Mastodon whether unpublished work entered into OpenAI products could have influenced the company’s model. OpenAI has not settled that question publicly.
For now, mathematicians are likely to do what they always do with a major claim: check the proof, argue about it, and test whether it holds. The $1 million prize is not awarded by press release. The real test will come from the mathematical community.
