🔍 Read the full analysis: A Question For OpenAI’s AI Mathematics: Where Do 722 Proofs Lead? on ThorstenMeyerAI.com
Get everyday essentials delivered free — and shop member deals
- Fast, free delivery on millions of items
- Access to Prime Big Deal Days deals on October 6–7
- Prime Video, Amazon Music and more included
TL;DR
OpenAI published 722 mathematical manuscripts produced by an unnamed, unreleased model, covering 372 families of results. The work includes claims about major open problems, but the claims have not been confirmed by outside mathematicians, and it is unclear whether the proofs will yield reusable ideas.
OpenAI published 722 mathematical manuscripts on Monday, generated by an unnamed model that the company has not released, presenting claims across fields including number theory, geometry and theoretical computer science. Some manuscripts claim results on famous open problems, but outside mathematicians have not yet confirmed them, leaving both their correctness and their potential value to other researchers unresolved.
The manuscripts are organized into 372 families of related results and were selected from work on roughly 4,000 problems posed to the model. OpenAI says the average result used about three hours of ChatGPT Pro reasoning compute. The collection was published under the Apache-2.0 license. OpenAI’s repository includes Lean formalizations for many, but not all, of the results; its README cautions that some results without formal proofs could have issues.
The catalogue includes claims concerning the Unique Games Conjecture, Hilbert’s tenth problem over the rationals, the isomorphism of nonabelian free group factors, and a zero-free region for the Riemann zeta function to the right of Re(s) = 11/12. It also includes claims involving the Hodge conjecture for CM abelian varieties and the Mahler conjectures in convex geometry. These are claims in the released manuscripts, not independently established breakthroughs.
OpenAI provided 10 abridged reasoning summaries for the 372 families. The source report says the company screened the roughly 4,000 problems for what it considered an appropriate level of significance, meaning the selection was made internally. Two manuscripts followed exceptions to the usual process: the Riemann write-up was edited by humans for readability, and the Hodge result was also treated differently. The available material does not fully detail that second exception.
722 proofs, one question: will any of OpenAI’s AI mathematics actually lead anywhere?
An unreleased, unnamed model produced claimed proofs of results that would each define a career. Sam Altman calls them “claims not yet confirmed by outside mathematicians.” The real question isn’t whether it’s impressive. It’s whether answers nobody understands become discoveries anyone can build on.
Same day: Alon, Bloom, Gowers, Litt, Sawin post a digested, human-verified version. The model for success.
Connes rigidity counterexample challenged within a day — constructed groups fail the required condition. Three rival machine “counterexamples” from different labs now circulate.
~10,000 agents, 88 hours, est. ~$22M at retail. Priority dispute; 25 Fields Medalists sign “A Severe Misalignment” — not saying it’s wrong, saying it’s not understood.
Altman now hedges at announcement — a shift from September. Verification has barely started.
Humans extract the technique, write it up, build on it. This is where downstream discovery comes from.
The question is answered; nobody learns anything reusable. Closes a door without opening a field.
The proof breaks, or proves a statement that doesn’t match the conjecture as mathematicians mean it.
The Unique Games Conjecture is the clearest case. Results like the optimality of Goemans–Williamson for Max-Cut are proved assuming UGC. A correct proof converts them all — no understanding required. A zero-free strip for zeta works the same way for prime-distribution results. Free group factors, Kadison, Mahler would redirect whole programmes — but how depends on the method, which means digestion.
Technology. A Navier–Stokes blow-up proof doesn’t change how anyone designs aircraft; engineering turbulence models never depended on the answer. Near-term consequences are mathematical, not industrial. “AI will cure cancer next” skips several steps.
“Verification abundance, adjudication scarcity” — making proof-checking cheap doesn’t reduce the burden of deciding what’s true and what matters. 722 manuscripts land on a review system built for a trickle, filtered by a selection nobody outside OpenAI made.
Humans re-deriving results, like Alon–Gowers et al. in May
Other people’s work building on these manuscripts
How many unformalized results survive expert checking
Do the Lean statements match the real conjectures?
Do any survive peer review?
Some of it, yes — where a literature is waiting (UGC), a correct proof pays off immediately; where a proof carries a new technique humans digest, it can open a field. Most of it, probably not on its own: at 722 manuscripts with 10 reasoning summaries, the Four Colour pattern is the likely default unless mathematicians are funded and given time. And some will be wrong — OpenAI says so itself. It’s an industry pattern, not one company’s: the forced-Euler result came from an Anthropic researcher, and rival machine-generated Connes “counterexamples” circulate from different labs. The proofs arrived this week. The discoveries, if they come, will arrive at the speed of human understanding.
From Proof Claims to Reusable Ideas
The importance of the release depends on more than whether individual statements are true. In mathematics, a proof can matter because it gives researchers a method they can adapt, not simply because it settles a question. Human understanding and independent checking will help determine whether these manuscripts become tools for further work or remain difficult-to-use answers.
The Unique Games claim illustrates the possible stakes. The conjecture is tied to a substantial body of theoretical computer science, including results about the limits of approximation algorithms. If the claim were correct and accepted, it could affect how researchers assess those results. But until mathematicians verify what has been proved and how, downstream consequences remain conditional.
The release also tests how AI-generated research enters a field whose standards depend on scrutiny, explanation and reuse. A large number of manuscripts does not by itself show that a field has advanced. The key measure will be what other mathematicians can check, understand and build on.
As an affiliate, we earn on qualifying purchases.
OpenAI’s Recent Math Results
This is described in the source report as OpenAI’s fourth major mathematics release this year. In May, its model produced a counterexample to the Erdős unit-distance conjecture. Five mathematicians then posted what they called a digested, human-verified version, illustrating how machine-generated work can be converted into a form the field can assess.
An August release called “Ten Advances” had mixed results. A claimed counterexample to Connes’s rigidity conjecture was challenged within a day; the critique argued that the constructed groups did not meet a condition required by the conjecture. The source report also notes that multiple machine-generated counterexamples to the same conjecture have circulated, underscoring the need to check whether a result addresses the exact mathematical statement at issue.
In September, OpenAI announced a Lean-formalized proof concerning finite-time blow-up in the Navier–Stokes equations, a Millennium Prize problem. The work was described as using about 10,000 agents over 88 hours. That announcement prompted a dispute over research priority and, three days later, a declaration signed by 25 Fields Medalists criticizing the use of famous problems as AI benchmarks without human understanding. The disagreement was about the purpose and practice of mathematical research, not a finding that the proof was wrong.
“Digested, human-verified version.”
— The five mathematicians who reviewed the Erdős unit-distance result
formal proof verification software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Independent Checks Still Pending
No outside confirmation is reported for the 722-manuscript collection as a whole. The source material does not establish which claims have been independently checked, which have been accepted by relevant experts, or whether formalized versions cover the most consequential arguments. OpenAI’s own warning about unformalized work adds a further qualification.
It is also unclear how the 372 families were selected beyond the company’s stated significance filter, and only 10 abridged reasoning summaries were supplied. The available information does not show how much of the reasoning a researcher would need to reconstruct or evaluate each result. Nor does it establish whether the proposed results will lead to new techniques or substantial follow-up research.
For any individual claim, the decisive questions are whether the proof matches the stated problem, whether its steps withstand expert review and whether other mathematicians can make use of its methods. Until those checks occur, the manuscripts should be treated as research claims awaiting evaluation, not as settled solutions.
As an affiliate, we earn on qualifying purchases.
What Mathematicians Must Verify
The next step is independent mathematical review: specialists will need to examine individual manuscripts, test formalizations where available and translate promising arguments into explanations the field can scrutinize. The earlier Erdős result offers one example of that process, but the source material does not say which of the newly released manuscripts are already being reviewed or when assessments may appear.
Readers should watch for corrections, detailed expert analyses, formal verification and follow-up papers that identify reusable methods. OpenAI has not provided a public timetable for those developments in the material supplied. The broader question—whether the work leads to new mathematics rather than merely proposed answers—will be answered gradually, result by result.
As an affiliate, we earn on qualifying purchases.
Key Questions
What did OpenAI release?
OpenAI published 722 mathematical manuscripts, organized into 372 families and generated by an unnamed, unreleased model. The work was selected from problems posed to the model across several areas of mathematics.
Have mathematicians verified the claimed breakthroughs?
The source report says the claims have not yet been confirmed by outside mathematicians. OpenAI’s repository also warns that some unformalized results could have issues.
Which major problems are included?
The manuscripts claim results involving the Unique Games Conjecture, Hilbert’s tenth problem over the rationals, free group factors, a region for zeros of the Riemann zeta function, the Hodge conjecture for CM abelian varieties and the Mahler conjectures. These remain claims pending review.
Why does it matter whether the proofs are understandable?
A correct proof can settle a question, but researchers often value a proof for methods that can be reused. Independent checking and human understanding will show whether the results can support further discoveries.
What happens next?
Mathematicians can review the manuscripts, examine available Lean formalizations and publish assessments or follow-up work. No review timetable for the collection is specified in the source material.
Source: ThorstenMeyerAI.com
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
