An internal OpenAI system has produced a mathematical proof and Lean formalization resolving the Navier–Stokes existence and smoothness problem, according to company announcements and independent reporting. The breakthrough, which tackles one of the seven Millennium Prize Problems established by the Clay Mathematics Institute in May 2000, demonstrates that fluid motion described by the Navier–Stokes equations can develop a singularity in finite time.
How Autonomous AI Agents Solved a Century-Old Fluid Dynamics Problem
The resolution emerged after a multiagent AI system operated on the problem for approximately 88 hours, launching on Tuesday, September 1, and reaching its conclusion on Saturday, September 5, according to OpenAI. Triggered by rumors that two Millennium Prize problems had been resolved, researchers deployed coordinating agents powered by an internal model significantly more capable than GPT-6 Astra. The agents utilized tools including code execution and cached internet access, sending 2.7 million messages and consuming roughly 130 billion output tokens during the Navier–Stokes effort alone.
According to OpenAI, the resulting proof establishes statements “C” and “D” in the official Millennium Prize formulation. The system proved that an initially smooth fluid at rest subjected to a smooth external force can develop a singularity while maintaining finite energy. The resulting dynamic forms an inward-spiraling, elongating vortex resembling spaghetti, where the central region shrinks and accelerates simultaneously.
Controversy and Attribution Surrounding the Mathematics Breakthrough
The announcement arrived amid accusations of complications and competitive tension involving independent human researchers. According to reports highlighted by Simon Willison and originating from Tristan Buckmaster, Buckmaster and Levent Alpöge—an accomplished mathematician working for Anthropic—had spent nearly a year investigating related problems using models like Claude and Codex (specifically GPT-5.6 Sol). After experiencing a breakthrough on August 15, the researchers encountered rumors that OpenAI was working on a related approach.
Buckmaster stated that OpenAI officials eventually confirmed their prompt was sent days after information concerning the NYU and Anthropic researchers’ work had circulated, though OpenAI maintained its agents did not access specific user data. While OpenAI offered a concurrent release and proposed recognizing the human researchers’ priority in a joint announcement, Buckmaster and Alpöge were not included as co-authors due to OpenAI’s relationship with Anthropic, according to published accounts.
Comparison of AI-Driven Mathematical Efforts
| Metric / Phase | Euler Regularity Problem | Navier–Stokes Resolution |
|---|---|---|
| Problem Scope | Unforced Euler equations (viscosity removed) | Navier–Stokes existence and smoothness |
| Agent Workforce | Nearly 100 agents | On the order of 10,000 concurrent agents |
| Time to Resolution | Approximately 50 hours | About 88 hours |
| Verification Method | Not specified in initial logs | 17 additional hours via GPT-6 Astra (Lean) |
Frequently Asked Questions
What is the Navier–Stokes existence and smoothness problem?
Named a Millennium Prize Problem by the Clay Mathematics Institute in 2000, it asks whether smooth solutions always exist for the Navier–Stokes equations governing fluid motion.
How did OpenAI’s system solve the problem?
According to OpenAI, a system of approximately 10,000 coordinating agents powered by an internal model generated an analytical proof and Lean formalization showing that fluid dynamics can develop a finite-time singularity.
Will OpenAI claim the Millennium Prize?
No. OpenAI stated that its goal in releasing the result is to report on the progress of its AI models rather than claim the cash prize.
What role did Lean play in the discovery?
Following the analytical proof, the Lean proof assistant was used to formalize and verify the findings, a process that took an additional 17 hours via GPT-6 Astra, according to OpenAI documentation.
Pro Tip: Stay updated on mathematical formalization tools like Lean as automated theorem proving increasingly intersects with frontier machine learning research.
Want to stay informed on developments in automated mathematics and artificial intelligence? Share your thoughts in the comments below or explore our archives for more breaking technology coverage.