OpenAI announced on Tuesday that its artificial intelligence technology has solved one of the Millennium Prize Problems, specifically the Navier-Stokes equations, which carry a $1 million reward for a solution. The company stated that a dedicated AI model, significantly more capable than its GPT-6 Astra, achieved this breakthrough in just 11 hours after an initial 50-hour effort on a related problem. OpenAI estimated the computational cost of this achievement at approximately $15 million, highlighting the substantial investment and advanced AI capabilities involved.
The announcement, however, was immediately met with controversy. Hours before OpenAI's declaration, Tristan Buckmaster of New York University and Levent Alpöge of Anthropic revealed their own AI-assisted progress on several related fluid dynamics problems, considered stepping stones to the Navier-Stokes solution. Buckmaster alleged that OpenAI had become aware of their work and then used an internal model to complete the Navier-Stokes problem, even claiming that OpenAI attempted to exclude Alpöge from authorship during a contentious phone call. OpenAI has categorically denied these allegations, stating that its model's proof is distinct and that no human or AI agents accessed Buckmaster and Alpöge's work in progress within its Codex system.
The dispute extends beyond credit to the very nature of the solution. Some mathematicians, including Fields Medalist Terence Tao, acknowledge the AI's ability to generate results quickly but express concern about the lack of human understanding and insight often accompanying these AI-generated proofs. The Clay Mathematics Institute, which administers the $1 million prize, has indicated that the evaluation process will be deliberately unhurried and rigorous, suggesting that the AI's solution may not be immediately recognized or fully accepted by the wider mathematical community without further scrutiny and clarity on its methodology.
Key figures in the dispute include Sébastien Bubeck, who leads OpenAI's math team, and Anthropic employee Levent Alpöge. The broader implication for the field of mathematics is a potential shift where AI can solve problems faster than humans can fully comprehend, raising questions about the future role of human intellect in mathematical discovery.
