TLDR
- OpenAI says roughly 10,000 AI agents solved part of the Navier-Stokes math problem in just 88 hours
- The agents exchanged 2.7 million messages and generated around 130 billion tokens during the process
- The internal model used is more powerful than any OpenAI has publicly released
- NYU mathematician Tristan Buckmaster says OpenAI may have used information from his unpublished research
- OpenAI says it won’t claim the $1 million prize and the solution has not been independently verified
OpenAI says an experimental system of around 10,000 AI agents solved part of the Navier-Stokes existence and smoothness problem in 88 hours. This is one of seven Millennium Prize Problems, each carrying a $1 million reward from the Clay Mathematics Institute.
We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics.
The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra.
The problem… pic.twitter.com/8zol3BPTL4
— OpenAI (@OpenAI) September 8, 2026
The Navier-Stokes equations describe how fluids like water and air move. The unsolved question asks whether a smooth fluid flow can eventually break down and produce speeds that grow without limit. OpenAI’s system says it can.
The proof describes a fluid vortex that stretches and concentrates until its speed blows up in a finite time, even while the system’s total energy stays finite.
The AI agents exchanged around 2.7 million messages and generated roughly 130 billion tokens while working on the problem. Based on OpenAI’s own pricing, that level of compute would cost around $10 million.
The model behind it is an internal system that OpenAI says is more powerful than GPT-6 Astra, its most advanced publicly available model. After the agents completed their work, Astra spent another 17 hours checking and formalizing the proof.
OpenAI said it started training this new model at the end of August. By September 1, the company had heard rumors that two Millennium Prize problems had been solved elsewhere. It then set its AI agents to work on the remaining problems.
By September 5, roughly 88 hours later, OpenAI had a proposed solution.
Mathematician Says OpenAI May Have Used His Unpublished Work
The announcement drew a fast response from Tristan Buckmaster, a math professor at New York University. He said he and Levent Alpöge, a mathematician at Anthropic, had been working on a closely related problem for about a year.
Buckmaster said the two had used OpenAI’s Codex tool during their research. He learned on September 3 that information about their progress had been passed to OpenAI.
He said OpenAI did not begin working on the Navier-Stokes problem until after that information reached them. He also pointed out that almost no other researchers were pursuing the same approach they were using.
OpenAI denied accessing any of Buckmaster and Alpöge’s work directly. The company said no user data was accessed and that it could not confirm any connection. However, it did say it could not fully rule out that data from their use of OpenAI products helped improve its models in some way.
OpenAI congratulated the two researchers on their work, calling it concurrent rather than prior. The company also noted that the two proofs differ and even prove slightly different results.
The solution has not been independently verified. The Clay Mathematics Institute requires proposed solutions to go through years of scrutiny before a prize is awarded.
OpenAI said it does not plan to claim the $1 million prize and described the result as a demonstration of how fast its research systems are improving.
Stop guessing and start investing with confidence. KnockoutStocks gives you the AI insights, market intelligence, and stock research you need to spot opportunities, cut through the noise, and make smarter investment decisions — all in one powerful platform.
Sign up today and get 50% OFF full access to our premium stock picks.
Simply use coupon code SPECIAL50 at checkout to claim your exclusive discount.







