‘Pure insanity’: Mathematicians will need years to make sense of OpenAI’s latest drop

OpenAI dropped nearly 400 AI-generated math results in 700-plus manuscripts, many still unverified.

OpenAI published nearly 400 AI-generated mathematical results spread across more than 700 manuscripts covering combinatorics, geometry, number theory, algebra, topology, probability, and mathematical physics. OpenAI said the claims are at different verification stages and that 300 of 719 top-line results, about 42 percent, had been formalized in Lean, with more formalizations promised. Mathematicians told The Verge the volume could take years to assess, and several warned that Lean proofs were inconsistent and did not always match the manuscript claims. Earlier OpenAI math write-ups were criticized for poor attribution, and researchers remain divided on whether this release improved on that record.