OpenAI Claims Breakthrough on Millennium Math Problem Amid Academic Backlash

OpenAI announced on Tuesday that an unreleased artificial intelligence system solved the Navier-Stokes problem, a complex fluid dynamics challenge that has remained unresolved by mathematicians for nearly nine decades. Using a swarm of approximately 10,000 AI agents, the internal model produced its proof in 88 hours. Although the achievement highlights rapid advancements in automated reasoning, the announcement sparked severe pushback from the academic community following accusations of unethical competitive practices and data misuse.

The fallout stems from the timing of OpenAI's release, which occurred one day after New York University mathematics professor Tristan Buckmaster published research on a closely related topic. Buckmaster co-authored the paper alongside Levent Alpöge, a researcher at competitor Anthropic who participated in an independent capacity. According to Buckmaster, after he contacted OpenAI to inquire about their timeline and training sets, an OpenAI researcher confronted him, asking, "Why would you ruin your career?" when Buckmaster indicated he would discuss the interaction publicly. The researcher subsequently added, "If you don’t want me to be nice, then I don’t have to be nice," while urging Buckmaster to publish the work with credit given to OpenAI's model and without Alpöge listed as a co-author.

Buckmaster also questioned whether OpenAI monitored his activity on its Codex coding platform while he worked on the problem. In response to allegations first reported by The Verge, OpenAI issued statements and a blog post explicitly denying the unauthorized use of private input. The company stated that "We (the researchers and the agents) did not see any of their work through any means until they released it publicly — in particular, no specific user data was accessed in order to solve this problem." However, OpenAI acknowledged it could not fully exclude the possibility that de-identified data derived from product usage influenced its models, though it emphasized that its final proof differed substantially from Buckmaster's.

Addressing how the project originated, OpenAI researcher Sébastien Bubeck stated during a press briefing reported by Science that the company decided to target a Millennium Prize problem after seeing rumors on X that outside scholars were nearing a breakthrough. Bubeck disputed portions of Buckmaster's account, specifically denying that he demanded Alpöge be removed as a co-author. Despite spending millions of dollars on compute resources for the last-minute push, OpenAI stated it has no intention of claiming the $1 million prize administered by the Clay Mathematics Institute, framing the initiative solely as a test of its model capabilities.

The aggressive pursuit of the proof has alarmed scholars who view OpenAI’s actions as a breach of long-standing academic norms. Abhishek Saha, a mathematics professor at Queen Mary University of London, noted that the company engaged in behavior standard mathematicians typically avoid, adding that traditional scooping is rare due to the deep specialization required. Matthew Ballard, associate director for scientific activities at the Institute for Computer-Aided Reasoning in Mathematics (ICARM) at the University of South Carolina, emphasized that mathematical progress depends on informal trust and the open sharing of unrefined concepts without fear of sudden commercial competition.

Other academics expressed deeper concerns regarding data security and corporate influence over fundamental science. Carnegie Mellon professor and ICARM director Jeremy Avigad warned that the possibility of AI systems extracting insights from user queries creates a chilling effect that could force researchers to become far more guarded. Brown University professor Brendan Hassett argued that tech firms must provide verifiable proof that user chat records remain private, citing the industry's history of utilizing copyrighted content. Meanwhile, Yang-Hui He, a fellow at the London Institute for Mathematical Sciences, cautioned that corporate secrecy risks reverting mathematics to an era of private patronage akin to the Renaissance.

While Oxford University professor Andras Juhasz characterized the milestone as a major promotional success for OpenAI, he questioned the long-term implications of deploying massive computational swarms against theoretical research. Scholars like Saha suggested that AI laboratories are unlikely to dedicate similar resources to everyday mathematical problems lacking high-profile publicity opportunities. Nevertheless, OpenAI's aggressive approach to outpacing academic rivals may have damaged its reputation among the professional research community it aimed to impress.