AI Is Rewriting Mathematics at a Pace Humans Can Barely Keep Up With

A Flood of Proofs, and a Field in Shock

In just one month, OpenAI has gone from shocking the world with a single landmark result to unleashing hundreds more — and the mathematics community is struggling to process both the scale and the speed. On Tuesday, October 6, OpenAI pushed a fresh batch of mathematical claims into public view, with an unreleased internal AI model producing findings touching 377 mathematical problems spanning algebra, number theory, theoretical computer science, mathematical logic, and topology. The model worked through roughly 4,000 mathematical problems during an internal evaluation period, with each attempted solution taking an average of about three hours of computing time.

Less than a month after the Navier-Stokes announcement, OpenAI went from one blockbuster claim to several hundred smaller ones, released all at once, with no peer review and no named model. The release is not a single paper — it is a pile of 722 manuscripts, organized by OpenAI into 372 result families, built on top of an internal system the company has not named and has no confirmed plan to ship. Among the most significant findings are results touching three of the remaining Millennium Prize Problems — including one involving the Riemann hypothesis, which concerns the distribution of prime numbers.

From Navier-Stokes to 377 More

OpenAI says it solved the 90-year-old Navier–Stokes mathematics problem in 88 hours, using 10,000 agents and a new artificial intelligence model. That proof shows that the dynamics of the Navier-Stokes equations for fluid motion can develop a singularity in finite time — specifically, that an initially smooth fluid at rest can develop a singularity in a finite time. The Navier-Stokes equations are used every day across physics and engineering, making the result far more than an abstract trophy.

Two mathematicians — Tristan Buckmaster of New York University and Levent Alpöge of Anthropic — had been working on a closely related problem using AI models, including OpenAI's. The relationship between the two advances, and whether OpenAI had access to the duo's progress, has been subject to intense speculation. The controversy strikes at a core trust question for AI-assisted science: whether researchers can safely use frontier labs' tools to work on unpublished discoveries.

An Advisory Board With No Real Power

In September, OpenAI announced a new independent advisory group hosted at the Institute for Advanced Study in Princeton, New Jersey — called the Advisory Group on Mathematics and Artificial Intelligence — meant to give mathematicians more input into the company's math-oriented research. These rapid breakthroughs had prompted 25 Fields Medal-winning mathematicians to publish an open letter warning that commercial AI laboratories risk compromising scholarly rigor by prioritizing competitive milestones over peer review. The advisory body, composed of nine prominent mathematicians, will evaluate the academic significance of emerging AI-generated proofs and coordinate their public release.

But critics say the board is largely symbolic. The group explicitly lacks authority to slow or redirect OpenAI's research trajectory, and the Institute for Advanced Study reaffirmed that the advisory panel holds no decision-making power over the company's operational pace or strategic direction. A newly formed group called the Association of Human Mathematics has gone further. Releasing more than 700 files at once, the AHM argues, is not a demonstration of scholarship but "a demonstration of power," and the group is calling on mathematicians to stop collaborating with OpenAI and return to a vision of science centered on human understanding.

What Comes Next for Math — and for Us

OpenAI has said it consulted the Advisory Group when deciding how to release the results, but critics note that group's own advisory statement began from the premise that frontier AI companies should not test advanced mathematical problems on internal models — a central point OpenAI has ignored. For 10 of the solutions, the company also provided a detailed account of how its AI model developed its solution — a gesture toward transparency that many in the field consider far too thin given the volume of claims being made.

The deeper question isn't just whether the proofs are correct. It's what happens to human mathematical culture if an opaque machine can outpace decades of scholarly effort in a matter of hours. Even OpenAI's own mathematicians have been taken aback by how fast the system has progressed, prompting serious internal conversations about how best to share these breakthroughs with the wider world. The math community now faces a reckoning that no advisory board, however prestigious, was designed to handle: how to remain relevant — and rigorous — in a field that AI is rewriting in real time.