OpenAI touts a Yang–Mills result, but a Millennium solution this year looks unlikely

Traders still see it as unlikely that OpenAI will announce a solution to a Millennium Prize problem by the end of the year, even after the company said its models had proved part of the Yang–Mills problem. Expectations edged up modestly since this story began and remain well short of even odds.

On Oct. 26, OpenAI announced that its models had produced a proof of the existence of pure Yang–Mills theory, one half of a problem posed by the Clay Mathematics Institute. The company said its models completed a formalization of that existence result on Oct. 16. The announcement likely helped nudge bets higher, but the Clay problem requires both existence and a positive mass gap, and OpenAI described only the first. A partial result does not count as a solution, so the announcement does not meet the bar for a full Yang–Mills solution.

The odds have been choppy. They hovered in a narrow range for most of the past two weeks, then jumped briefly on Oct. 25 before plunging sharply on Oct. 26 as the news landed and traders weighed what OpenAI had actually claimed. They recovered through the day on Oct. 27, with another brief burst late in the evening, and finished only slightly above where they started. Over the longer window, traders remain split on whether OpenAI announces a Millennium solution by the end of 2027.

The swings follow a volatile month. On Oct. 6, OpenAI released 722 mathematical manuscripts, including a result on a weaker, "quasi-Riemann" statement rather than the Riemann Hypothesis itself. The next day, Retraction Watch reported, OpenAI withdrew three manuscripts after errors were found. Mathematicians also raised concerns that the results had not been adequately vetted.

OpenAI said an outside expert group will shepherd the physics work and write it up, and that it is holding back the AI-produced materials in the meantime. No publication or review date has been announced. Only a public announcement that OpenAI or its researchers had also established the mass gap, or had solved another qualifying problem, would change the picture. Navier-Stokes does not qualify, and neither do announcements that credit only outside researchers.