OpenAI pulls three AI-generated math papers one day after release

What OpenAI Released and Pulled

OpenAI released 722 AI-generated mathematics papers on October 6, 2026, which contained 372 result families produced by an unnamed proprietary model. The papers aimed at resolving around 4,000 unresolved problems in mathematics and claimed advances on numerous longstanding conjectures, including progress on several Millennium Prize Problems.

The Verification Problem

After the initial release, 300 of the 719 results had been formally verified in Lean, a proof assistant that translates a proof into formal code and checks it step by step. Critics pointed to the proprietary nature of the model, which outside researchers cannot inspect or reproduce its process.

The Release and Its Impact

The release was informed by the Advisory Group on Mathematics and Artificial Intelligence, known as AGMAI. This is not OpenAI's first clash with the math community, as the company previously withdrew a $1M sponsorship from a Caltech AI math contest after criticism from 771 mathematicians.

The Sign Error and Its Consequences

The flawed proof sat in algebraic geometry and contained a sign error, which means a plus where a minus should be, or the reverse. This error can turn a valid chain of logic into one that proves nothing at all, and every argument built on top of it falls with it. The corrected catalog now stands at 719 papers, with the withdrawn papers replaced by revised versions.