OpenAI Claims Solution to Millennium Prize Problem Amid Plagiarism Allegations
The AI lab says it solved the 200-year-old Navier-Stokes equations, but mathematicians question the origin of the method.
OpenAI has claimed a historic breakthrough in mathematics, asserting that an internal AI model solved the Navier-Stokes Millennium Prize Problem. The announcement has ignited a fierce debate over scientific integrity and the opacity of AI development cycles.
According to OpenAI, a model significantly more capable than GPT-6 Astra proved that the Navier-Stokes equations—which describe fluid dynamics—can "blow up," meaning fluid speed becomes infinite in finite time. The company reports the solution was reached in 88 hours using approximately 10,000 autonomous AI agents. To provide academic rigor, OpenAI released a formalization of the proof using the Lean programming language, a tool designed to verify mathematical logic.
The Fluid Dynamics Puzzle
The Navier-Stokes equations have stumped researchers for two centuries. Solving the "existence and smoothness" problem is one of seven Millennium Prize Problems, each carrying a $1 million reward. The challenge lies in determining whether solutions to these equations always exist and remain smooth, or if they can develop singularities. OpenAI's claim that the equations can develop a singularity represents a definitive, if controversial, answer to a foundational question in physics.
Allegations of Intellectual Theft
The breakthrough has been met with skepticism by some in the academic community. Mathematician Tristan Buckmaster, who has been working on related fluid-dynamics research with colleague Levent Alpöge, alleged that OpenAI's research direction mirrored their own work on the Euler equations. Specifically, Buckmaster raised concerns that OpenAI adopted a "forcing" method after becoming aware of his and Alpöge's progress, questioning whether private Codex data was utilized to guide the AI.
OpenAI has denied these allegations. Sébastien Bubeck, a mathematician at OpenAI, stated, "We did not use their prompt or proofs to prompt our models." While the company denied using specific user data, it acknowledged that the effort was launched after rumors circulated that other labs, including Anthropic, might have solved a major mathematical problem.
A Paradigm Shift in Discovery
This event signals a potential shift where AI evolves from a calculation tool into a generator of original mathematical proofs. However, the friction between OpenAI and independent researchers highlights a growing tension. While AI labs operate with rapid, opaque development cycles, the scientific community relies on transparent, peer-reviewed traditions. As Diego Córdoba noted, if the solution is indeed verified, "that will be a big surprise for us."
What Remains Unconfirmed
While the Lean certificates provide a layer of formal verification, the broader mathematical community has yet to reach a consensus on the proof's validity. The industry is now watching to see if other independent teams can replicate the result or if the controversy over the "arms race" for mathematical discovery will lead to new standards for AI-generated research attribution.