OpenAI takes aim at another Millennium Prize

OpenAI appears to be treating the Millennium Prize Problems like a personal checklist.

According to The Information, sources say the company is close to solving the Hodge Conjecture. It marks the second Millennium Prize math challenge OpenAI has taken on.

Internal staff expect a solution relatively soon. The Hodge Conjecture basically asks whether complex geometric shapes defined by polynomial equations can be broken down and described using simpler algebraic building blocks.

Walking on eggshells around mathematicians

To tackle the earlier Navier-Stokes problem, OpenAI deployed a variant of its upcoming model codenamed "Doug."

That previous announcement, however, turned into what The Verge described as a PR disaster with the broader math community. Between that backlash and recent model control issues, OpenAI is now pumping the brakes on a victory lap.

The company is reportedly trying to figure out how to work alongside mathematicians before announcing anything, hoping to avoid another public feud.

Why it matters

If frontier models can legitimately crack abstract milestones like the Hodge Conjecture, it proves machine reasoning is reaching into the highest echelons of theoretical science. The problem-solving chops inside the "Doug" variant suggest AI is moving far beyond basic pattern matching.

Still, OpenAI's biggest challenge right now isn't the math—it's the optics. Pure mathematics demands painstaking peer review and bulletproof formal verification, not Silicon Valley hype cycles.

We think slowing down and planning a collaborative rollout is the right move. After burning bridges during Navier-Stokes, rushing another grand claim would only draw deeper skepticism. If OpenAI wants real credibility in academia, it has to play by math's ground rules.