# OpenAI says agents produced Navier-Stokes Millennium Prize solution > The company said around 10,000 agents running a model significantly more capable than GPT-6 Astra produced the proof in about 88 hours. Clay still labeled the problem unsolved on September 9. - Publication: Go Big News - By Go Big News Staff (Staff News) - Category: AI - Published: 2026-09-09T15:30:56.17+00:00 - Canonical: https://www.gobignews.com/ai/openai-says-agents-produced-navier-stokes-millennium-prize-solution --- [OpenAI](https://openai.com) said on September 8 that a group of AI agents using an internal next-generation model significantly more capable than GPT-6 Astra had produced a solution to the Navier-Stokes existence and smoothness problem. The Clay Mathematics Institute named it one of seven Millennium Prize Problems in 2000. The company announced the result in a post on X and a research writeup that day. It released a paper titled “Finite Time Blowup for Navier-Stokes” and Lean 4 formalizations. OpenAI said the agents reached the result on September 5, about 88 hours after the first agents were launched. The group involved on the order of 10,000 concurrent agents. Formalization and verification in Lean took another 17 hours via GPT-6 Astra. Go Big News [reported on September 4](https://www.gobignews.com/ai/openai-starts-gpt-6-astra-rollout) that OpenAI had started rolling out GPT-6 Astra. The Navier-Stokes equations describe how fluids move by treating a fluid as a continuous medium rather than tracking individual molecules. OpenAI said the question, unresolved for about 90 years, asks whether smooth three-dimensional fluid motion can break down. It defined a singularity as fluid speeds growing without bound in finite time despite viscosity. The company said its system produced an analytical proof that an initially smooth fluid at rest, subject to a smooth force and with finite energy throughout, can develop a singularity in finite time. OpenAI said this establishes statements C and D in the official prize formulation, covering whole-space R^3 and the periodic torus. It described the flow as a vortex that spirals inward and becomes elongated. OpenAI said the Navier-Stokes agents sent 2.7 million messages and used about 130 billion output tokens. Across all attempted problems, the company said, agents sent 4.9 million messages and used about 300 billion output tokens. OpenAI said it has been training the internal model since August 28 and that training is ongoing. It began the evaluation on September 1 after hearing rumors that two Millennium Prize problems had been resolved. The company later tied the rumors to Levent Alpöge, an Anthropic employee, and Tristan Buckmaster, a math professor at NYU, and recognized their priority on the forced Euler problem. OpenAI said its researchers and agents did not see Alpöge and Buckmaster’s work until its public release and that no specific user data was accessed to solve the problem. It said it cannot rule out that de-identified product-usage data helped improve its models. As of September 9, Clay still labeled the Navier-Stokes problem unsolved. The institute allocates $1 million to each Millennium Prize Problem. Its rules require publication in a qualifying outlet, a wait of at least two years, general acceptance in the global mathematics community as determined by Clay, and an answer to the official problem description. OpenAI said it does not intend to claim the prize. --- ## Companies in this story - Openai: https://www.gobignews.com/company/openai --- ## X profiles in this story - @openai: https://www.gobignews.com/profile/openai