OpenAI AI Solves Navier-Stokes Millennium Prize Problem

OpenAI has announced that an internal AI model has solved the Navier-Stokes existence and smoothness problem, resolving one of the seven Millennium Prize Problems. The breakthrough demonstrates that smooth three-dimensional fluid motion can develop a singularity in finite time.

To reach the solution, OpenAI deployed a multi-agent system consisting of approximately 10,000 concurrent agents. Working over a period of roughly 88 hours, the system generated an analytical proof and a Lean formalization. For the Navier-Stokes problem specifically, the process consumed approximately 130 billion output tokens and 2.7 million messages. The internal model utilized for this task is described by OpenAI as being significantly more capable than GPT-6 Astra. The resulting proof resolves statements 'C' and 'D' of the official problem formulation, confirming that speeds in a 3D incompressible fluid can grow without bound.

A Century of Mathematical Mystery

The Navier-Stokes existence and smoothness problem has remained unresolved for approximately 90 years. It centers on whether the continuum approximation of fluid dynamics—the mathematical foundation used for everything from aircraft design to weather forecasting—can fundamentally break down. In 2000, the Clay Mathematics Institute designated it as one of seven Millennium Prize Problems, each carrying a $1 million reward. Until this announcement, only one of these problems, the Poincaré conjecture, had been solved.

Implications for AI and Science

This achievement marks a potential paradigm shift in the capacity of artificial intelligence to perform high-level mathematical reasoning and autonomous scientific discovery. By coordinating thousands of agents to tackle a problem that eluded human mathematicians for nearly a century, OpenAI has demonstrated a new scale of synthetic reasoning. Despite the magnitude of the discovery, OpenAI stated it does not intend to claim the $1 million prize from the Clay Mathematics Institute.

Emerging Controversies

While the technical feat is historic, it has sparked a dispute regarding intellectual property and data privacy. NYU mathematician Tristan Buckmaster and Anthropic researcher Levent Alpöge have questioned whether the model leveraged their own unpublished research on related fluid-dynamics problems to find the solution. OpenAI has denied these allegations. Mark Chen, OpenAI's Chief Research Officer, stated that "no people or AI systems searched through user data to solve this problem."