OpenAI AI solves Navier‑Stokes problem in 88 hours, sparks credit dispute

Artificial intelligence has crossed a formidable threshold in mathematical research, though the achievement is instantly mired in acrimony over ownership and methodology. According to Digitaljournal, researchers announced that an unreleased artificial intelligence model solved a notoriously difficult physical and mathematical puzzle that has confounded human experts for generations. The Navier-Stokes problem, which describes how fluids such as air and water move, serves as a cornerstone for disciplines ranging from weather forecasting to aircraft design. Yet what was billed as a landmark scientific triumph has ignited an intense credit dispute between industry labs and university mathematicians.

The Navier-Stokes equations, first formulated in the mid-19th century, present a profound paradox in classical physics. While they use Newton's second law of motion to explain fluid dynamics reliably in everyday applications, mathematicians have long struggled to prove whether their solutions remain smooth or inevitably break down over time into a singularity—a theoretical point where speeds grow infinitely. Designated in 2000 by the Clay Mathematics Institute as one of seven Millennium Prize Problems, the challenge carries a reward of one million US dollars for a correct solution. Only one of the original seven problems has been resolved since the list's inception.

Media additions

OpenAI researcher Sebastien Bubeck stated that the computational costs ran emphatically into the millions of dollars, representing roughly a thousandfold increase over previous mathematical experiments. Across all the problems attempted, agents sent millions of messages and consumed billions of output tokens.

However, the celebration was cut short by New York University mathematician Tristan B. Buckmaster. Working alongside Levent Alpöge, a mathematician at rival lab Anthropic, Buckmaster had spent months pursuing a narrow, unconventional approach involving a smooth force. As detailed by English, Buckmaster alleged that news of his progress reached OpenAI in early September 2026, prompting the firm to deploy massive resources along the exact same technical route. Buckmaster further claimed that interactions with OpenAI’s coding tools during his team's research might have allowed the underlying model to absorb their specific reasoning.

OpenAI vigorously rejected any suggestion of academic impropriety. In briefings reported by Aa, representatives maintained they had no prior access to the rival papers and initiated their work independently. While acknowledging in secondary statements that it could not entirely rule out that de-identified usage data had helped train underlying models, OpenAI insisted its final proofs differed from those of the human researchers. Furthermore, the company declared it would not claim the monetary bounty from the Clay Mathematics Institute. Professor Martin Bridson, President of the Clay Mathematics Institute, told AFP that the formal evaluation process requires publication in a peer-reviewed journal and two years of scrutiny by the mathematical community before any committee is convened.

The controversy underscores broader business coverage surrounding intellectual property and generative tools. Creators and major news publishers have raised persistent legal challenges regarding how tech firms scrape human works. Meanwhile, eminent figures like Fields Medalist Terence Tao have praised the underlying mathematics achieved by the human-AI collaborations while warning that unreadable automated proofs could complicate peer review. As the mathematical community digests the new proofs, attention will turn to whether independent experts can verify the formal Lean code and how academic institutions adapt to machine-driven discovery.