OpenAI's claimed Navier-Stokes solution ignites feud with mathematicians

OpenAI says ~10,000 agents and tens of millions in compute produced a Navier-Stokes Millennium Prize solution in 88 hours; 25 Fields Medalists signed an open letter warning of attribution and plagiarism harms, and NYU's Tristan Buckmaster accuses OpenAI of…

OpenAI claims an advanced unreleased model, run with roughly 10,000 agents and tens of millions of dollars of compute over 88 hours (described by mathematicians as a marathon weekend), produced a solution to the Navier-Stokes problem — one of seven $1 million Millennium Prize problems set by the Clay Mathematics Institute in 2000, of which only the Poincaré conjecture had previously been solved. The proof remains unverified. NYU professor Tristan Buckmaster, who was pursuing the problem with Anthropic researcher Levent Alpöge, accused OpenAI of racing his team, possibly benefiting from his Codex prompts, and pressuring him not to credit his Anthropic-employed collaborator; OpenAI categorically denied racing the team and the Codex claim. OpenAI researcher Sébastien Bubeck acknowledged offering Buckmaster compute and authorship arrangements, which Buckmaster characterized as a bribe. Twenty-five Fields Medal-winning mathematicians signed an open letter — following the June Leiden Declaration on LLM proofs — arguing rushed AI proofs raise severe attribution and plagiarism questions and could destroy the culture of open research, with many mathematicians describing the field as shaken by OpenAI's competitive conduct. OpenAI also withdrew sponsorship of a Caltech math event after researcher criticism, and mathematicians fear their Codex usage may be fed into OpenAI's new models.

• OpenAI says an advanced unreleased model, run with ~10,000 agents and tens of millions of dollars of compute over 88 hours, produced a Navier-Stokes solution; the proof remains unverified.
• Navier-Stokes is one of seven $1 million Millennium Prize problems set by the Clay Mathematics Institute in 2000; only the Poincaré conjecture had been solved before this claim.
• NYU professor Tristan Buckmaster, working with Anthropic researcher Levent Alpöge, accused OpenAI of racing his team, possibly benefiting from his Codex prompts, and pressuring him over crediting his Anthropic-employed collaborator; OpenAI…
• OpenAI researcher Sébastien Bubeck acknowledged offering Buckmaster compute and authorship arrangements, which Buckmaster called a bribe.
• 25 Fields Medal-winning mathematicians signed an open letter citing attribution, plagiarism, and threats to the culture of open research; it follows the June Leiden Declaration on LLM proofs.
• OpenAI withdrew sponsorship of a Caltech math event after researcher criticism.
• Mathematicians fear their Codex usage may be fed into OpenAI's new models, and more than a dozen mathematicians described the field as shaken by OpenAI's conduct.