OpenAI says it has made progress on a second Millennium Prize problem

OpenAI says its agents are working on a second of the seven Millennium Prize problems.

The statement

In a statement to The New York Times, OpenAI said: “In addition, since the completion of Navier-Stokes, we have made substantial progress on another Millennium Prize problem. We are working through how to share these results thoughtfully.”

The statement was given on September 9, 2026, the day after OpenAI announced its Navier-Stokes result. OpenAI has not said which problem it means. Online speculation points to the Hodge conjecture. OpenAI has not confirmed that.

What is and is not known

“Substantial progress” is not a claim of a solution. The statement gives no timeline, no problem name and no description of the method. It tells readers only that the company believes it has something worth announcing and has not yet decided how.

The context is a contested week. OpenAI’s Navier-Stokes result has not been accepted by the Clay Mathematics Institute, which awards the prizes, and it drew a dispute with NYU mathematician Tristan Buckmaster over prior work. Mathematicians have also objected to AI-sponsored events in their field.

Why the wording matters

The phrase “share these results thoughtfully” suggests OpenAI expects scrutiny over process as much as over the mathematics itself: how the result was reached, how it was checked, and how credit is assigned to any human work it builds on. Those are the questions the Navier-Stokes announcement raised, and they will apply again to whatever OpenAI publishes next.