A vortex stretches like spaghetti in the Navier-Stokes proof that OpenAI says its AI agents produced. The company announced the claimed solution on September 8, targeting a longstanding question about three-dimensional fluid motion. OpenAI says the swirling fluid develops a singularity (a breakdown in the mathematical description) within finite time. The supplied reporting does not establish independent verification. What exactly does the company claim its agents have found? [1, 2]
Navier-Stokes Proof Targets Fluid Breakdown
OpenAI describes a fluid that begins smoothly, then spirals inward and becomes increasingly elongated. Its researchers say the equations eventually assign it infinite speed within a finite interval. Ven Chandrasekaran, an OpenAI computer scientist, described that outcome at a press briefing reported by Nature. Real liquids and gases cannot behave that way, he said. The claim concerns the limits of the equations as a description of physical reality. [1, 2]
Mathematicians have struggled with the underlying question for roughly 90 years, according to OpenAI. Can the equations preserve smooth three-dimensional motion, or can that mathematical description break down? The Clay Mathematics Institute included the question among its Millennium Problems, whose solutions carry a $1 million prize. OpenAI says its proposed example takes the breakdown route. The company claims a solution; the supplied sources do not establish that mathematicians have independently verified it. [1, 2]
Engineers already use these equations to model fluids, from airflow around aircraft wings to blood moving through arteries. New Scientist describes those applications while reporting on a separate mathematical advance. Readers may recognize the broader territory from our coverage of methods for tackling partial differential equations. Here, the question narrows to whether a particular fluid model can develop a mathematical failure. OpenAI says it pushed thousands of agents toward that question. [1, 3, 5]
Ten Thousand Agents, One Problem
OpenAI says its internal model group reached the Navier-Stokes proof in 88 hours, using around 10,000 coordinating AI agents. The company attributes the work to a next-generation model that it describes as significantly more capable than GPT-6 Astra. It also says training continues. Those performance descriptions come from OpenAI itself. The material available here supplies no independent benchmark assessment of that comparison. [1]
OpenAI mathematician Sebastian Bubeck gave Nature a more detailed account of the escalating effort. Researchers initially tested their prototype on all six unsolved Millennium Problems. They concentrated on fluid motion on September 1, after hearing reports about work by Levent Alpöge and Tristan Buckmaster. Bubeck said 1,000 agents first answered a simplified version in 50 hours. The company then increased the effort to 10,000 agents for the full Navier-Stokes problem. Nature reports those figures as the researchers’ account. It does not provide a separate audit of the run. Nor does its account explain how the 50-hour stage relates to the 88 hours in OpenAI’s public post. [1, 2]
The agents’ output, OpenAI says, includes an analytical argument and a Lean formalization. Can a computer-checkable proof settle the claim? [1]
What Lean Adds to Checking
Mathematicians use Lean to express mathematical arguments in code that computers can check for logical errors. New Scientist explains that process in its coverage of Alpöge and Buckmaster’s work. OpenAI likewise says it produced a Lean formalization alongside its analytical proof. A claim that a formal proof exists still needs scrutiny of the actual mathematical result. The source material here reports the announcement without supplying an independent examination of OpenAI’s files. [1, 3]
OpenAI links the announcement to a page titled On the Navier–Stokes Millennium Prize Problem. The research package could not retrieve that page’s full text. Its short feed summary therefore cannot support additional claims about assumptions, verification results or the proof’s contents. The accessible X announcement supplies the company’s central claims. Nature supplies press-briefing details and the surrounding research chronology. Neither accessible account documents a completed independent review of the Navier-Stokes proof. [1, 2, 4]
Clay Mathematics Institute president Martin Bridson told Nature that the announcements marked an exciting day for mathematics. The quoted response welcomes the prospect of major advances; it does not announce a prize award. The supplied reporting also does not establish peer-reviewed publication of OpenAI’s argument. Readers therefore have an announced result to examine, with its verification status still unresolved in these sources. Meanwhile, other researchers have reached nearby equations by a different route. [2]
Human Teams Reach Nearby Equations
Tristan Buckmaster and Levent Alpöge announced three results involving close relatives of Navier-Stokes, according to New Scientist. Their work addresses the Boussinesq approximation and Euler equations, with substantial help from language models. Two results accompanied Lean formalizations; the pair held back another while completing its formalization. New Scientist describes steps toward the full problem, rather than treating those published results as a general Navier-Stokes solution. [3]
Nature reports that the pair released a paper on September 7 concerning fluids with zero viscosity (resistance to flowing). They used Anthropic’s Claude and OpenAI’s Codex and Astra models. Nature also reports their statement that a more general result would follow. Separately, Anima Anandkumar and collaborators released a zero-viscosity result using a physics-informed neural network, rather than a general-purpose language model. These accounts describe distinct teams, methods and scopes. They do not independently confirm OpenAI’s claim. [2]
Terence Tao expressed optimism about extending Alpöge and Buckmaster’s methods, according to New Scientist. Camilla Nobili, at the University of Surrey, also considered the direction promising. But she expected Navier-Stokes to require more than repeating the same technique. Viscosity introduces effects that tend to smooth fluid motion, complicating the search for a breakdown. How would a proposed singularity survive those smoothing effects? That question separates the nearby results from the larger target. [3]
What the Claim Leaves Open
David Silvester, at the University of Manchester, told New Scientist that the related results could serve as stepping stones. He also cautioned against expecting immediate practical gains. Fluid models already work well in engineering, he said, so a mathematical resolution need not change existing applications. His comments concern the research discussed in that report. They offer context for the larger question without evaluating OpenAI’s Navier-Stokes proof. [3]
OpenAI emphasizes the resources it committed and the capabilities it attributes to its developing model. Buckmaster, in New Scientist’s account, emphasizes AI’s growing ability to amplify human mathematical effort. Both accounts put AI inside the research process, but they assign different roles to people and models. The available reporting does not establish an independently verified winner or a prize recipient. The underlying arguments, their assumptions and their scope remain the material readers need to judge. [1, 2, 3]
Readers following the Navier-Stokes proof can start with OpenAI’s announcement and compare it with Nature’s reporting. The next substantive question concerns the proof itself: can independent mathematicians confirm that its argument establishes the claimed finite-time breakdown for the stated problem? The supplied reports do not answer that question. They leave readers with OpenAI’s assertion and several related advances whose precise mathematical scope matters. [1, 2, 3]
- WEBSITE OpenAI [@OpenAI]. (2026, September 8). We’re sharing a solution to the Navier-Stokes Millennium Prize Problem [Post]. X. [Article Link]
- ONLINE NEWS Castelvecchi, D. (2026, September 8). OpenAI claims huge maths breakthrough on a famed ‘Millennium Problem’. Nature. [Article Link]
- ONLINE NEWS Sparkes, M. (2026, September 8). Major breakthrough made on famous Millennium maths problem. New Scientist. [Article Link]
- WEBSITE OpenAI. (n.d.). On the Navier–Stokes Millennium Prize Problem. [Full page unavailable in the research package; listed for reference, not as evidence of proof verification.] [Article Link]
- ONLINE NEWS PerEXP Teamworks. (n.d.). Innovative method offers effective resolution for a multitude of applications involving partial differential equations. [Related coverage.] [Article Link]
