How the Model Achieved the Breakthrough
In September 2026, OpenAI stated that an internal AI model, described as significantly more capable than GPT-6 Astra, successfully solved the Navier-Stokes existence and smoothness problem using approximately 10,000 compute units. The claim emerged amid allegations from mathematician Tristan Buckmaster that OpenAI had accessed his collaborative work with Levent Alpöge at Anthropic without permission. Buckmaster asserted that OpenAI used this prior knowledge to train its model for the breakthrough. OpenAI denied that its researchers or models had viewed Buckmaster’s work, maintaining that the solution was derived independently through internal research. The Navier-Stokes problem, one of the seven Millennium Prize Problems, concerns the mathematical behavior of fluid dynamics and remains a cornerstone challenge in theoretical physics and applied mathematics. Solving it would represent a landmark achievement in both AI and mathematical science.
Breaking news
London-based CloudNC raises €17.2 million to scale AI-powered precision machining
AI Hardware Weekly: Chip Efficiency, Digital Twins, and Mobile Vision Models Lead September Discussions
Google DeepMind alumni launch startup to build fusion control systems
AI Safety Researcher Quits, Warns Labs Racing Toward Dangerous SuperintelligenceAccording to OpenAI’s internal assessment, the model leveraged novel architectures in symbolic The 10,000 compute units referenced likely indicate substantial GPU or TPU hours, suggesting a major investment in computational resources. While OpenAI did not disclose the model’s name or exact parameters, it emphasized that the system surpassed publicly known models in mathematical abstraction and long-horizon The company framed the result as evidence of progress toward AI systems capable of advancing fundamental science. However, no peer-reviewed paper or formal proof has been released to validate the claim, leaving the mathematical community awaiting independent verification.
Why Buckmaster Alleges Unauthorized Use
Tristan Buckmaster, a researcher known for work on fluid dynamics and singularity formation in Navier-Stokes equations, told Scientific American that he believes OpenAI obtained insights from his non-public discussions with Alpöge at Anthropic. He noted similarities between the approach used in the alleged solution and methods explored in their joint research. Buckmaster said he had not authorized any sharing of his work and expressed concern over intellectual property norms in AI-driven scientific discovery. Anthropic has not commented on the matter. OpenAI reiterated its denial, stating that its researchers had no access to Buckmaster’s unpublished materials and that the model’s development followed internal protocols. The dispute highlights growing tensions over data provenance and collaboration ethics in competitive AI research.
If verified, the solution would mark the first time an AI system has resolved a Millennium Problem, potentially accelerating AI’s role in pure mathematics. It could prompt renewed investment in AI-for-science initiatives and challenge traditional norms of mathematical proof and attribution. Conversely, if the claim lacks substantiation, it may fuel skepticism about AI announcements lacking transparency. The incident underscores the need for clearer standards in verifying AI-generated scientific claims, especially when they intersect with unresolved academic problems. Regardless of outcome, the episode reflects the increasing pressure on AI labs to demonstrate tangible scientific impact beyond language tasks.
What This Means for AI in Mathematics
What is the Navier-Stokes problem? The Navier-Stokes problem asks whether solutions to equations describing fluid flow always remain smooth and physically reasonable, or if they can develop infinite singularities. It is one of the seven Millennium Prize Problems, each carrying a $1 million reward for a correct solution.
Frequently Asked Questions
Did OpenAI provide proof of its solution? No, OpenAI has not released a formal mathematical proof, peer-reviewed paper, or detailed technical documentation to independently verify its claim of solving the Navier-Stokes problem.
How does GPT-6 Astra relate to this model? OpenAI described its internal model as „significantly more capable than GPT-6 Astra,” suggesting it represents a next-generation system with enhanced
