OpenAI reports an agent-produced finite-time blowup for Navier–Stokes
On 8 September 2026 OpenAI published a claimed resolution of the Navier–Stokes existence and smoothness problem, one of the Clay Mathematics Institute's Millennium Prize Problems, announcing that an internal model not available to the public, run as roughly 10,000 concurrent agents, had constructed a finite-time singularity for the three-dimensional incompressible equations with a smooth forcing term. OpenAI said the effort began on 1 September and that the model took about 88 hours; it released a manuscript together with a formalization in the Lean proof assistant, and said it would not claim the $1 million prize. On 11 September the Clay Mathematics Institute said the problem had "apparently been settled" and that its process for evaluating the work and assigning credit would be deliberately unhurried, with no prize awarded and no verification issued.
Why It Mattered
This is the first time a Millennium Prize Problem has been addressed by a machine-driven proof effort, and the first case in which a frontier lab has asserted a result at that level of difficulty. Two features distinguish it from earlier claims of AI mathematical achievement. The first is verification: the argument was formalized in Lean, which gives mathematicians a mechanical check on correctness independent of whether they can follow the human-readable write-up, and which is what allowed the Clay Institute to respond within three days rather than after years of refereeing. The second is scale: the work was done by a swarm of agents operating in parallel over days, a mode of mathematical production with no prior analogue, and one that shifts the binding constraint on certain proofs from insight to compute budget. The result also arrived inside a dispute that will likely outlast the mathematics. OpenAI began work on 1 September after hearing a rumour of progress elsewhere, and the rumour traced to Levent Alpöge, an Anthropic employee, and Tristan Buckmaster of NYU, who announced a closely related forced-Euler blowup on 7 September. OpenAI said neither its researchers nor its agents saw that work before public release, but conceded it could not rule out that de-identified data derived from the pair's use of its products had helped improve its models. That admission opened a question with consequences well beyond fluid dynamics: whether researchers can use commercial AI tools on unpublished work without the tool provider gaining an advantage. Charles Fefferman, who wrote the Clay problem statement, has said credit belongs primarily to the earlier human work of Diego Córdoba and Luis Martínez-Zoroa. Substantive gaps remain — the construction uses a forcing term, so the unforced global regularity question is untouched — and independent scrutiny is ongoing.
Who Built It
OpenAI
Applications
- Mathematics
- Fluid Dynamics
- Automated Theorem Proving