🔍 Read the full analysis: OpenAI’s 722 AI Mathematics Proofs Prompt A Question About What’s Next on ThorstenMeyerAI.com
Get smart everyday buys 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, covering 372 families of results drawn from roughly 4,000 problems. The company says some claims concern major open problems, but external mathematicians have not confirmed them, and many manuscripts lack formal verification. Whether the work produces reusable mathematical ideas remains unknown.
OpenAI published 722 mathematical manuscripts on Monday, presenting results generated by an unnamed, unreleased model across 372 families of related work. The catalogue includes claims about long-standing open problems, but OpenAI has said the results have not been confirmed by outside mathematicians, leaving their accuracy and potential impact unresolved.
The manuscripts span number theory, geometry, operator algebras, topology, theoretical computer science and mathematical physics. OpenAI says the work was selected from roughly 4,000 problems posed to the model; results took an average of about three hours of ChatGPT Pro thinking compute. The collection is published under the Apache-2.0 license.
Among the claims are a proof of the Unique Games Conjecture, a resolution of Hilbert’s tenth problem over the rationals, and results concerning nonabelian free group factors, the Riemann zeta function and the Hodge conjecture for CM abelian varieties. These are claims in manuscripts, not findings independently established by the mathematical community.
OpenAI’s repository includes Lean formalizations for many, but not all, results. Its README warns that some unformalized results could have issues. The company also provided ten abridged reasoning summaries from the 372 families. According to the source account, the Riemann write-up was edited by humans for readability, and the Riemann and Hodge results followed exceptions to the usual process.
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.
Verification Will Shape the Impact
The release’s significance depends less on the number of manuscripts than on what survives expert scrutiny and whether mathematicians can use the reasoning. A formally checked proof can establish that a particular argument follows within a chosen formal system, but it does not by itself show that a result is important, that the formalization matches the intended mathematical claim, or that the work offers ideas people can build on.
That distinction matters especially for a claim such as the Unique Games Conjecture. If it were proved in a form accepted by specialists, it could affect a substantial body of theoretical computer science that relies on the conjecture when analyzing the limits of approximation algorithms. But the manuscript’s appearance in OpenAI’s catalogue does not establish that outcome. Independent review and clear communication of the argument are still needed.
There is also a question about how mathematical progress is measured. A proof may settle a famous problem yet contribute few reusable techniques; another may offer methods that support further work. The source account contrasts this with OpenAI’s May report of a counterexample to the Erdős unit-distance conjecture: mathematicians produced a digested, human-verified version the same day. That example illustrates one possible path from machine output to work the field can evaluate.
As an affiliate, we earn on qualifying purchases.
OpenAI’s Earlier Math Claims
The 722 manuscripts are described as OpenAI’s fourth major mathematics release this year. Earlier releases offer relevant context, though they do not determine whether the new claims are sound.
In May, OpenAI reported that its model had found a counterexample to a 1946 Erdős conjecture. Five mathematicians then published what they called a digested, human-verified version. In August, OpenAI announced ten advances; one claimed counterexample involving Connes’s rigidity conjecture was challenged within a day. The criticism said the constructed groups did not meet a condition required by the conjecture.
In September, OpenAI announced a Lean-formalized proof concerning finite-time blow-up for the Navier–Stokes equations, generated, according to the company, by about 10,000 concurrent agents over 88 hours. That announcement prompted a dispute over priority and the purpose of using famous problems as benchmarks. The source account says 25 Fields Medalists signed a declaration criticizing the approach as misaligned with mathematical practice; their objection was about the goals and consequences of the work, not a finding that the proof was wrong.
formal verification tools for mathematicians
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Which Manuscripts Will Hold Up
Independent verification remains the central unknown. The source material does not report outside mathematicians confirming the major claims in this new catalogue. It also does not say which of the 372 families have been examined by specialists, whether any purported proof has been formally checked end to end, or whether the formalizations accurately capture each intended statement.
The selection process is another limitation: OpenAI chose which results to publish from roughly 4,000 problems, and the account says nobody outside the company made that selection. The ten abridged summaries offer only a small view of the full set. It is not yet clear how much of the underlying reasoning other researchers can readily inspect, or how long specialist review will take.
Finally, the broader consequences are unknown. Some results may fail review, some may prove correct without generating useful methods, and others could be developed into new work by mathematicians. The release alone cannot establish which outcome applies to any particular manuscript.
mathematical proof assistant software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Independent Review Comes Next
The next meaningful milestone is scrutiny by researchers outside OpenAI: checking the statements, examining proof steps and formalizations, and determining whether the results add something reusable. Where a claim is correct but difficult to follow, mathematicians may need to produce clearer accounts, as they did with the earlier Erdős result.
OpenAI has not, in the source material, provided a timetable for external validation or identified which manuscripts it expects specialists to review first. Readers should treat the headline claims as unconfirmed until researchers publish assessments. Over time, the more consequential test will be whether verified arguments lead to follow-on work, not simply how many manuscripts the model produced.
As an affiliate, we earn on qualifying purchases.
Key Questions
What did OpenAI publish?
OpenAI published 722 mathematical manuscripts grouped into 372 families, attributed to an unnamed, unreleased model. The company says they were selected from roughly 4,000 problems posed to the model.
Have mathematicians confirmed the results?
Not according to the source material. OpenAI described the results as claims that have not yet been confirmed by outside mathematicians, and its repository warns that some unformalized results could have issues.
What is the Unique Games Conjecture claim?
One manuscript claims a proof of the Unique Games Conjecture, a major open problem in theoretical computer science. The catalogue’s inclusion of the manuscript does not mean the claim has been independently verified.
Does a Lean formalization prove a result is important?
A Lean formalization can help check that a proof follows within the formal system, but it does not on its own establish the result’s broader significance or whether its methods will help mathematicians make further discoveries.
What happens next?
Researchers outside OpenAI will need to examine the manuscripts and any available formalizations. The source material gives no review timetable, so it remains unclear when the major claims will receive independent assessments.
Source: ThorstenMeyerAI.com
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
