🔍 Read the full analysis: OpenAI’s AI Mathematics Has Built 722 Proofs—What Happens Next? on ThorstenMeyerAI.com
Get office and shipping supplies 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 attributed to an unnamed, unreleased model, grouped into 372 families and selected from roughly 4,000 problems. The work includes claims about major open problems, but outside mathematicians have not yet confirmed the results; whether the proofs yield reusable ideas remains open.
OpenAI published 722 mathematical manuscripts on Monday, presenting work by an unnamed, unreleased model across fields including number theory, geometry and theoretical computer science. The collection includes claims about major open problems, but the results have not been confirmed by outside mathematicians, leaving their correctness and potential impact unsettled.
The manuscripts are organized into 372 families of related results and were selected from roughly 4,000 problems posed to the model, according to OpenAI’s post and repository. The source account says each result took about three hours of ChatGPT Pro thinking compute on average. OpenAI says it filtered the problems for an “appropriate level of significance”; the selection was made internally.
The claims span a striking range: they include a proof of the Unique Games Conjecture, a proposed resolution of Hilbert’s tenth problem over the rationals, and a result on nonabelian free group factors. Other manuscripts claim a zero-free region for the Riemann zeta function to the right of Re(s) = 11/12, the Hodge conjecture for CM abelian varieties, and results related to the Mahler conjectures. These are claims in the released manuscripts, not independently established breakthroughs.
OpenAI published ten abridged reasoning summaries, rather than summaries for all 372 families. Many results have Lean formalizations, but not all. The repository warns that “some of the unformalized results could have issues.” The source account also identifies two exceptions to the usual process: the Riemann write-up was edited by humans for readability, and the Hodge result followed a different procedure.
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.
Can These Proofs Change Mathematics?
The collection’s importance will depend on more than whether individual statements are true. In mathematics, a proof can matter because its techniques become tools other researchers can apply. A correct result that no one can interpret or build on may settle a question without changing how the field works.
The source account contrasts the release with OpenAI’s earlier work on the Erdős unit-distance conjecture. After the model produced a counterexample, five mathematicians published a digested version they had checked. That process gave other researchers a form they could evaluate. By contrast, the source describes a claimed counterexample involving Connes’s rigidity conjecture that was disputed within a day over whether the constructed groups met the conjecture’s requirements.
Those examples point to the practical test for the new catalogue: whether mathematicians can verify the arguments, explain their ideas and find uses for them. Claims involving the Unique Games Conjecture, for instance, could matter to theoretical computer science if confirmed, because many results about approximation algorithms are proved under that conjecture. But neither a claimed proof nor its possible consequences should be treated as established before review.
As an affiliate, we earn on qualifying purchases.
OpenAI’s Earlier Math Releases
The 722-manuscript release is described in the source account as OpenAI’s fourth major mathematics release this year. In May, a model produced a counterexample to the Erdős unit-distance conjecture, a claim dating to 1946. Five mathematicians then posted a human-verified, more accessible account of the result—a useful example of how machine-generated work can enter mathematical discussion.
In August, OpenAI announced ten advances. One claimed counterexample, concerning Connes’s rigidity conjecture, faced a rapid critique that said the constructed groups did not satisfy the condition the conjecture required. In September, OpenAI announced a Lean-formalized proof that the Navier–Stokes equations can blow up in finite time, generated using about 10,000 concurrent agents over 88 hours, according to the source material.
The Navier–Stokes announcement also prompted a dispute over research priorities. Three days later, 25 Fields Medalists signed a declaration titled “A Severe Misalignment of AI in Mathematics.” The source says their concern was that using famous problems as benchmarks without human understanding could work against the aims of mathematics; it does not characterize their complaint as proof that the result was wrong.
““A Severe Misalignment of AI in Mathematics.””
— The declaration signed by 25 Fields Medalists
Verification and Value Remain Open
No independent confirmation of the catalogue’s headline claims is established in the supplied source material. It is also unclear how many manuscripts will withstand expert scrutiny, how long review will take, and whether the Lean formalizations cover the central arguments or only parts of them. OpenAI’s warning about unformalized results makes that distinction relevant.
Even if a result is correct, its lasting contribution is not yet known. Mathematicians may extract a reusable technique, accept a proof that closes a problem without creating new methods, or find a flaw or mismatch between the statement proved and the conjecture researchers intended. The source does not provide independent evaluations for each of the 372 families or explain how OpenAI ranked the selected problems.
How Mathematicians Will Test the Work
The immediate next step is independent mathematical review. Researchers will need to examine the manuscripts, verify their statements and proofs, and determine what the formalizations establish. For results without formal verification, that review may involve reconstructing arguments and checking technical details directly.
OpenAI has released the manuscripts and repository, but the supplied material does not state a timetable for outside assessments or identify which results will receive priority. The clearest evidence of progress will be a result that experts can verify and explain in a form others can use—not the number of manuscripts published. Until that happens, the catalogue is a large set of consequential claims, rather than a confirmed set of mathematical breakthroughs.
Key Questions
What did OpenAI release?
OpenAI published 722 mathematical manuscripts attributed to an unnamed, unreleased model. They are grouped into 372 families and were selected from roughly 4,000 problems, according to the source material.
Have mathematicians verified the claimed proofs?
The supplied source does not establish independent confirmation of the catalogue’s major claims. OpenAI’s repository also cautions that some unformalized results could have issues.
What major problems do the manuscripts address?
The claims include results concerning the Unique Games Conjecture, Hilbert’s tenth problem over the rationals, nonabelian free group factors, the Riemann zeta function, the Hodge conjecture for CM abelian varieties and the Mahler conjectures.
Why does a correct proof not automatically transform mathematics?
A proof may settle a question without providing methods other researchers can reuse. Its broader value depends on whether mathematicians can understand the argument, extract new ideas and apply them elsewhere.
What happens next?
Researchers must review the manuscripts, check the proofs and assess what any formal verification covers. The source material gives no timetable for those reviews or for independent judgments on each result.
Source: ThorstenMeyerAI.com
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
