The world of computational mathematics is heating up once again. OpenAI announced that its internal AI model proved the Navier-Stokes equations can reach a “blow up” point - a condition where mathematics predicts infinite velocity. Navier-Stokes is one of the seven Millennium Prize Problems, carrying a $1 million reward from the Clay Mathematics Institute for anyone who solves it.
OpenAI claims the proof process took 88 hours and involved roughly 10,000 AI agents working concurrently. They utilized a next-generation model more advanced than GPT-6 Astra that has not yet been released to the public. The result is a 100-page proof document verified using Lean software.
However, the 88-hour figure does not tell the whole story. A fierce dispute over discovery credit unfolded in its wake.
Three Days to the Announcement
New York University mathematician Tristan Buckmaster, alongside Anthropic researcher Levent Alpöge, claimed to have found the same solution on August 22, 2026, after working with AI models for nearly a year. On September 3, Buckmaster said he informed an OpenAI mathematician about his project, emphasizing that the conversation was personal and not affiliated with any company.
Three days later, Sébastien Bubeck from OpenAI revealed that their internal model had already produced a 100-page proof for Navier-Stokes. The “forced Navier-Stokes” approach used was nearly identical to Buckmaster’s route.
Tensions escalated when Buckmaster claimed Bubeck gave him two narrow options. The first option: OpenAI would publish its findings one day after Buckmaster’s team. The second option: Buckmaster would publish the research on his own without Alpöge’s name, as Alpöge is an employee at Anthropic, OpenAI’s rival. Buckmaster accused Bubeck of making threats, allegedly asking, “Why would you ruin your career?” and saying, “If you don’t want me to be nice, then I don’t have to be nice.”
Rebuttals and an Unfinished Race
The allegations were immediately dismissed by the company. Sam Altman defended his colleague, stating that Bubeck acted with high integrity throughout the process. Bubeck countered that his team only later learned Buckmaster had only proven the Euler equations rather than the full Navier-Stokes, which led to a change in the joint release plan.
Buckmaster published his research anyway, listing Alpöge as a co-author. OpenAI offered no acknowledgment to the two researchers other than a single reply tweet. Despite the controversy, Fields Medalist Terence Tao described Buckmaster and Alpöge’s mathematical work as a notable achievement.
This marks the second major claim in a matter of weeks, following Anthropic’s announcement that its Claude model proved Fermat’s Last Theorem. To date, neither OpenAI’s 100-page proof nor Buckmaster’s solution approaching the requirements for the $1 million prize has received independent verification.
The AI race has expanded from text generation to pure mathematics, but the corporate scramble for the spotlight shows no sign of cooling down. Via Decrypt.
Read also: Bitcoin Breaks $80,000 Ahead of Fed Meeting - But 162,000 New Jobs Data Brings Different Scenario
Disclaimer: This article is for informational and educational purposes only, not financial advice. Cryptocurrency assets are highly volatile and carry significant risk. Always do your own research (DYOR) and never invest more than you can afford to lose.
