Key Takeaways
OpenAI said on September 8, 2026 that an internal AI model produced a proof, formalized in the Lean proof assistant, that the Navier-Stokes equations can form a singularity in finite time.
The system used on the order of 10,000 coordinating AI agents, which reached the result in about 88 hours while sending 2.7 million messages and using roughly 130 billion output tokens.
The Navier-Stokes problem is one of seven Clay Mathematics Institute Millennium Prize Problems, and independent mathematicians have not yet verified the claim.
OpenAI said on September 8, 2026 that an internal AI system produced a proof that the Navier-Stokes equations, which describe how fluids move, can develop a singularity in finite time. The company released a written proof and a formalization in Lean, a proof assistant that machine-checks every step, and posted the code on GitHub.
The Navier-Stokes existence and smoothness question has stayed open for roughly 90 years. In 2000 the Clay Mathematics Institute named it one of seven Millennium Prize Problems, each carrying a $1 million reward. OpenAI said it would not claim the prize, releasing the work instead to show how fast its models are moving.
We do not intend to claim the Millennium Prize for this result.
OpenAI, in its research announcement
The result came from a system of coordinating agents running on a model OpenAI describes as more capable than its GPT-6 Astra release. The group that resolved Navier-Stokes involved on the order of 10,000 concurrent agents, and it reached the answer on September 5, about 88 hours after the effort began. Lean verification took another 17 hours. On that one problem the agents sent 2.7 million messages and used roughly 130 billion output tokens.
The proof describes a vortex that spirals inward and stretches like spaghetti, speeding up until its velocity grows without bound while its total energy stays finite. Independent mathematicians have not yet confirmed the work, and researchers at Anthropic and New York University released a related result on the Euler equations the same week.
The most notable part of this is the method. A landmark proof did not come from one mind but from thousands of agents pooling partial insights, the kind of machine coordination now reaching fields WYDE covers, from the onchain infrastructure growing on Base to the legal framework Wyoming built for nonprofit DAOs. Peer review is the next test. Worth watching whether it holds.
People Also Ask
It is one of seven problems named by the Clay Mathematics Institute in 2000, asking whether smooth three-dimensional fluid flows can break down into a singularity. Each carries a $1 million prize.
OpenAI says an internal AI model produced a proof, verified in Lean, that a singularity can form in finite time. Independent mathematicians have not yet confirmed it, and OpenAI is not claiming the prize.
How did OpenAI's AI agents produce the proof?
Roughly 10,000 coordinating agents explored different approaches, and their insights were merged using OpenAI's Codex tool, reaching a result in about 88 hours followed by 17 hours of Lean verification.
Why does an AI-generated math proof matter?
It suggests AI systems can now attack open research problems directly rather than only assist, though the result still needs human peer review before it stands.
Sources
OpenAI, On the Navier-Stokes Millennium Prize Problem. Clay Mathematics Institute, Millennium Prize Problems. Unite.AI, coverage of the announcement.