> ## Content Index
> Fetch the complete content index at: https://www.tao.media/llms.txt
> Use this file to discover other available public pages before exploring further.

# OpenAI Says Its AI Agents Solved the Navier-Stokes Millennium Prize Problem
- URL: https://www.tao.media/openai-says-its-ai-agents-solved-the-navier-stokes-millennium-prize-problem/
- Published: 2026-09-08T18:56:39.000Z
- Updated: 2026-09-08T18:56:39.000Z
- Description: OpenAI released a proof and Lean formalization while framing the result as evidence of faster AI research progress.
- Author: Bart Hillerich
- Tags: OpenAI, News, AI

[OpenAI](https://www.tao.media/tag/openai/) says an internal system of AI agents has produced a solution to the Navier-Stokes existence and smoothness problem, one of the seven Millennium Prize Problems and one of the best-known open questions in mathematical physics.

The company announced the result in an [X post](https://x.com/OpenAI/status/2097374640582668336?s=20&ref=tao.media) and a detailed [research publication](https://openai.com/index/navier-stokes-solution/?ref=tao.media), where it released both a written proof and a Lean formalization. OpenAI says the proof shows that the Navier-Stokes equations for three-dimensional fluid motion can develop a singularity in finite time, meaning the mathematical description of a smooth fluid flow can break down.

Three things make the announcement notable. If the mathematics community accepts it, the proof would resolve a problem that has stayed open for roughly 90 years. It also offers one of the clearest public examples yet of a frontier AI system coordinating a large-scale proof search across thousands of agents, rather than just assisting research. Moreover, it's a window into what's next after OpenAI's GPT-6 Astra model, which recently had [Jensen Huang claiming that "AGI has arrived."](https://www.tao.media/jensen-huang-says-agi-has-arrived-after-openais-gpt-6-astra-launch/)

## What OpenAI Says It Proved

The Navier-Stokes equations describe how fluids move. They are used across fields such as aerodynamics, weather modeling, and blood-flow analysis, and they treat a fluid as a continuous medium rather than tracking every molecule individually.

The Millennium Prize version of the problem asks whether smooth three-dimensional incompressible fluid motion always remains smooth, or whether the equations can produce a finite-time singularity even when the initial conditions are well behaved. In practical terms, the question is whether the model can force speeds inside the fluid to grow without bound in a finite amount of time.

OpenAI says its system found the latter: a construction in which an initially smooth fluid at rest develops a singularity. According to the company's write-up, the fluid is acted on by a smooth force, maintains finite energy throughout the process, and still forms a breakdown in finite time.

![](https://storage.ghost.io/c/78/0b/780ba906-b1a7-4bf0-873c-bdd5c32e5331/content/images/2026/09/image-9.png)

"The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra."

The proposed solution is based on a vortex structure. OpenAI describes it as a spinning swirl of fluid that spirals inward and stretches out, with a central region shrinking and accelerating while preserving finite energy. 

The technical challenge, the company says, is making the terms in the equations become large while canceling precisely enough that the external force remains smooth. A valid counterexample cannot simply insert an infinite force by hand. The breakdown has to arise from the dynamics of the equations themselves.

## How the Agent System Found the Proof

OpenAI says the proof was produced by an internal model "significantly more capable than GPT-6 Astra," with training still underway. The company describes the result as part of a broader evaluation effort launched after it heard rumors that other Millennium Prize-related work may have been completed.

According to the research publication, OpenAI began the effort on Sept. 1 and assigned groups of agents to open Millennium Prize problems and related mathematical questions. The system included agents with access to tools such as cached internet reading and code execution, and the agents were organized into communicating groups of varying sizes.

The group that produced the Navier-Stokes result involved about 10,000 concurrent agents. OpenAI says the agents arrived at the resolution in roughly 88 hours, with Lean formalization and verification taking an additional 17 hours through GPT-6 Astra.

Across all attempted problems, OpenAI says its agents sent 4.9 million messages and used about 300 billion output tokens. For the Navier-Stokes work specifically, the agents sent 2.7 million messages and used about 130 billion output tokens.

The company says it also tested the system on a related question for the Euler equations, which remove the viscosity term from Navier-Stokes. OpenAI says a smaller group of nearly 100 agents worked for about 50 hours to produce an unforced Euler regularity disproof, which then helped guide the later Navier-Stokes work.

## Why Lean Formalization Matters

OpenAI's release includes a link to a Lean-formalized proof, which is important because mathematical claims of this size typically require extensive expert review before they are broadly accepted.

Lean is a proof assistant that allows mathematical arguments to be expressed in a form that can be checked by software. A Lean formalization does not automatically make a result socially or academically settled, but it can reduce ambiguity by forcing the proof into a machine-checkable structure.

That is especially relevant here because the claimed result is both highly technical and unusually consequential. The Clay Mathematics Institute's Millennium Prize Problems were designed to recognize some of the deepest unsolved questions in mathematics. A proof or disproof of Navier-Stokes smoothness would normally undergo sustained scrutiny by specialists before being treated as resolved.

OpenAI appears to recognize that distinction. The company says it is releasing the result to report on AI progress and does not intend to claim the Millennium Prize.

## Why Solving Navier-Stokes is important

On assessing the implications of this discovery, Jubal Hardin published the [following message](https://x.com/Jubal%5FHardin/status/2097358710087590009?s=20&ref=tao.media) on X:

> "What makes solving Navier-Stokes important? Fluid dynamics affects more than you realize in your everyday life. This breakthrough — if it is true — gives you: Cheaper, safer flying and driving. Air over a wing or around a car is a Navier-Stokes problem. Better control of that flow means less drag, so planes and trucks burn less fuel, stay quieter, and stall less often. Weather you can actually plan around. Storms, jet streams, and ocean currents are giant fluid problems. Sharper turbulence physics is one of the remaining limits on hurricane tracks, heat waves, and flood forecasts. That is the difference between “maybe evacuate” and a clear warning. Hearts, stents, and artificial valves. Blood is a fluid. Doctors already simulate flow through arteries and implants from CT/MRI scans. A more reliable theory of swirling, high-speed blood flow would help design stents, heart valves, and aneurysm treatments that work better. Cleaner power from wind and turbines. Wind hitting a turbine blade, steam inside a power plant, and fuel mixing in an engine are all the same equations. Better models mean more electricity from the same wind farm and engines that waste less fuel Ships, pipes, and dirty air. Oil in a pipeline, water in city mains, smoke over a city, and waves slamming a hull all follow Navier-Stokes. Better predictions cut leaks, pump energy, and help track pollution after a spill or wildfire. Many other everyday effects: buildings, sports, and rain. The same math decides whether a skyscraper sways in a gale, why a dimpled golf ball flies farther, how air moves through a house’s vents, and even how raindrops form in clouds. Those are not abstract puzzles; they are why planes fly, forecasts exist, and a stadium roof doesn’t peel off. None of this happens instantly because of a math problem solution by AI. But everything that does happen is going to help engineer a better life for all of us. This is big."

## The Broader AI Research Signal

The announcement is also a statement about AI capability; OpenAI's internal systems are clearly entering a new phase of research acceleration, particularly in mathematics.

The notable part is how OpenAI organized a large population of agents around a hard research objective, let them explore different variants and approaches, and then cross-pollinated useful intermediate results. The company says Codex was used to consolidate insights from agent groups before follow-up prompts guided additional work.

That resembles a research operation more than a single chatbot response. It combines frontier model capability, tool use, parallel exploration, formal verification and orchestration across thousands of agents. If the Navier-Stokes proof withstands review, the result would become a tremendous achievement of AI systems contributing directly to frontier mathematics.

It also raises questions about how scientific credit, verification and disclosure should work when AI systems produce major research outputs. OpenAI notes that its effort began after rumors involving Levent Alpöge, an Anthropic employee, and Tristan Buckmaster, a math professor at NYU. The company says it later learned their work concerned the forced Euler problem, not the same Navier-Stokes result, and says it recognizes their priority on that separate achievement.

OpenAI also says its researchers and agents did not see that work before it was released publicly. The company adds that while it cannot rule out the possibility that de-identified data derived from product usage helped improve its models, it says the proofs and precise Euler results differ significantly.