OpenAI published a large batch of mathematical manuscripts produced by an unreleased internal frontier model in a GitHub repository, openai/math, under the title "Sharing AI progress in mathematics." The first release contained 722 manuscripts organized into 372 "families" (a principal result together with companion arguments, consequences, or alternative proofs), including the Kakeya conjecture (the maximal conjecture in three dimensions and the Hausdorff-dimension conjecture in four), a "quasi-Riemann hypothesis" stating that every Dirichlet L-function is zero-free for Re s > 7/8, the Hodge conjecture for CM abelian varieties, and the Birch–Swinnerton-Dyer formula for elliptic curves of Selmer corank zero or one. OpenAI said the vast majority of results came from the same procedure, that the model was posed approximately 4,000 problems, and that each result used on average about three hours of ChatGPT Pro thinking compute. Exceptions were the zero-free region for the Riemann zeta function and the Hodge work, and the Re(s) > 11/12 writeup was human edited for readability. "Many, but not all" manuscripts had Lean formalizations, and OpenAI acknowledged that unformalized results "could have issues." It said it had consulted the Advisory Group on Mathematics and Artificial Intelligence hosted at the IAS, whose formation it announced on September 21, and followed the group's public recommendations by releasing 10 reasoning summaries, compute estimates, and statistics on attempted problems; it also said it would fund workshops and conferences on understanding major AI-produced results. The New York Times reported the release as findings on "377 math problems."
OpenAI releases 722 manuscripts from an unreleased model covering 372 result families, including Kakeya in four dimensions
Participants
Also mentioned
Not parties to this event, but named in the text above.
Tags
- research
- mathematics
- frontier-models
Sources
- Sharing AI progress in mathematicsen
- Sharing AI progress in mathematics (Wayback Machine)en
- openai/math (GitHub)en
- openai/math README at the initial commit (October 6, 2026)en
- OpenAI Releases Findings on 377 Math Problems, Further Roiling Fielden
- OpenAI dumps 372 AI-generated math proofs on GitHub, telling the academic world to keep upen
Later developments
7 October 2026 · Development
OpenAI updated the repository, withdrawing three manuscripts, including the rational Hodge conjecture for products of K3 surfaces, after a sign error in "Algebraicity of Weil classes on split abelian eightfolds" invalidated a construction used by two dependent papers, and revising 14 others. It put the share of top-line results formalized at 300 of 719, about 42%. The same day the Association for Human Mathematics issued a statement saying "Releasing over 700 files at once is not a demonstration of scholarship, but a demonstration of power," urging mathematicians to discontinue work with OpenAI and noting that the advisory group from which OpenAI "has claimed to derive its legitimacy" had opened its initial statement by saying frontier AI corporations should not test advanced mathematical problems on internal models.
10 October 2026 · Development
The math blog Proofs and Prompts collected more than a hundred responses from mathematicians, summarized by The Decoder. Peter Scholze urged patience, writing that "mathematics is a marathon, not a sprint," and raised the possibility that such systems could break widely used encryption. Terence Tao wrote that the AI-generated proofs appeared to "introduce clever new ideas that will be fruitful once digested" but that he was "deeply frustrated" by the absence of people who could discuss or teach the work. Hugo Duminil-Copin, a 2022 Fields Medalist, wrote: "The shock is immense. I am paralysed."
26 October 2026 · Development
OpenAI added a manuscript titled "Local smoothing for the Euclidean wave equation," with Lean formalizations, to the repository. It claims a proof of Sogge's local smoothing conjecture in every dimension two and above, and thereby of the Bochner–Riesz conjecture, the restriction conjecture for the sphere, and the Kakeya conjecture in all dimensions. The same update added a manuscript titled "The Exponential Time Hypothesis is False." As of mid-November no specialist had published an assessment.