OpenAI claims Navier–Stokes breakthrough, faces credit dispute

The announcement that OpenAI's agents solved the Navier–Stokes problem would normally be a landmark achievement in mathematics. However, the credit dispute with Tristan Buckmaster and Levent Alpöge has shifted the focus to how AI-driven research may bypass traditional academic norms. Both teams used an approach pioneered by Diego Córdoba and Luis Martínez-Zoroa, and while independent arrival at the same method is possible, the coincidence has fueled suspicions.

The deeper issue is the changing nature of mathematical progress. Buckmaster and Alpöge, working with publicly available models for nearly a year, made significant progress but did not achieve a full solution. OpenAI, in contrast, solved the full problem in a few days using an internal model and about 10,000 concurrent agents at a cost of millions of dollars. This disparity raises questions about what role individual human mathematicians can play when such resources are concentrated in a few companies.

The outcome of this controversy could shape how AI companies handle credit and transparency in research. If OpenAI's models did use Buckmaster and Alpöge's work without acknowledging it, that would reflect poorly on the company. But even if they did not, the episode suggests that solving major mathematical problems may increasingly depend on access to frontier AI capabilities, potentially leaving most human mathematicians without the resources to compete.