Analysis
OpenAI said it has made "substantial progress" on a second Millennium Prize math problem, according to The Information, weeks after the company announced on September 8 that an internal model had solved the Navier-Stokes problem -- a 200-year-old open question about the mathematics of fluid motion -- using as many as 10,000 AI agents working concurrently over 88 hours, a claim CNN and the Washington Post both covered at the time.
Rumor Versus Confirmed Statement
Secondary commentary on social media has claimed OpenAI is "very close" to resolving the Hodge conjecture specifically -- one of the seven original Millennium Prize problems, each carrying a $1 million prize from the Clay Mathematics Institute for a verified solution. That specific claim is not confirmed by OpenAI's own statement, which is deliberately vaguer: "substantial progress on another Millennium Prize problem," without naming which one. Conflating the Hodge-conjecture rumor with OpenAI's actual, less specific disclosure overstates what the company has confirmed.
“Conflating the Hodge-conjecture rumor with OpenAI's actual, less specific disclosure overstates what the company has confirmed.”
The Scale Behind The First Claim
What's more verifiable is the mechanics behind the Navier-Stokes claim: OpenAI's internal model reportedly ran as many as 10,000 AI agents concurrently over 88 hours to reach its result -- an orchestration scale that is itself a significant engineering achievement independent of whether the underlying mathematical proof holds up. Coordinating that many agents on a single unified proof, rather than 10,000 independent attempts, implies substantial infrastructure for agent-to-agent coordination and result verification that OpenAI hasn't detailed publicly.
Why Peer Review Matters More Than The Announcement
Millennium Prize problems carry their reputation precisely because they have resisted the world's best human mathematicians for decades; a claimed AI solution requires the same rigorous, often years-long peer-review process any human-derived proof would need before the mathematics community treats it as verified, let alone before the Clay Mathematics Institute would consider awarding its associated prize. That review process for the Navier-Stokes claim was still ongoing as of this reporting, roughly ten days after the initial announcement -- nowhere near long enough for a claim of this magnitude to be considered settled by mathematical community standards.
Two Unverified Claims, One Review Queue
A second unverified claim emerging before the first is fully peer-reviewed compounds rather than resolves the evaluation challenge: mathematicians now have two OpenAI claims to independently verify instead of one, on a timeline the company's own announcement cadence is outpacing. If either claim holds up under full scrutiny, it would be a genuinely historic AI-research milestone; if the Navier-Stokes proof develops holes under closer review, it would cast a longer shadow over how much weight to put on the newer, still-unnamed second claim.