🔍 Read the full analysis: A Mathematician Says Claude Helped With Riemann. Here’s What Was Checked on ThorstenMeyerAI.com
Get smart everyday buys delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Anthropic mathematician Levent Alpöge says he and Claude found proofs of the Riemann Hypothesis and the stronger Generalized Riemann Hypothesis, releasing a Lean code archive rather than a readable proof. Early checks found the code compiled, but a key comparator test was still pending, and mathematicians have not accepted the result.
Mathematician Levent Alpöge, who works at Anthropic, said in a note posted November 20 that he believes he and the company’s Claude AI helped prove the Riemann Hypothesis and the broader Generalized Riemann Hypothesis. He released a short explanation and a large archive of Lean proof code, but the claim has not been accepted by mathematicians, and a key independent check was still pending as of November 21.
Alpöge’s note says he “believe[s]” he discovered the proof while working with Claude. It asserts that the same method extends to the Generalized Riemann Hypothesis, which concerns a wider family of mathematical functions. The note is not a complete readable proof: it gives a sketch, a timeline of the work and links to the formalized argument in 14 parts.
According to computer scientist Scott Aaronson, the archive contains about 17.4 million lines of Lean code. Lean is a proof assistant that checks whether formalized steps follow logically from stated assumptions. Alpöge says a human-readable version would exceed 250,000 pages and chose not to publish it. The note does not identify which Claude model was used.
MIT mathematician Andrew Sutherland told Scientific American that the code compiled independently “with only the three standard axioms.” That is an early technical check, not confirmation that the argument proves the intended theorem. Kevin Buzzard said a “comparator run” was still awaited. The source account describes that test as checking whether the formal code proves the real mathematical statement rather than a different or incorrectly encoded one.
Explainer · Riemann hypothesis claim
From claim to accepted proof: six checks, and where this one stands
What the hypothesis says, and the path every proof claim has to take before mathematicians accept it.
The claim in one picture
Riemann’s guess (1859): every non-obvious zero of the zeta function lies on this one line. If true, the primes are spread out as evenly as possible.
The checks a proof claim must pass
- 1Claim postedA 4-page note plus Lean code, November 20, 2026✓ done
- 2Code compiles elsewhereIndependent compile reported by an MIT mathematician✓ reported
- 3Comparator runRight statement? Only standard axioms? Several Lean kernels?… pendingWe are here · November 21, 2026
- 4Humans understand itA readable proof that experts can follow○ open
- 5Journal publicationA refereed mathematics journal of worldwide repute○ open
- 6Two years + general acceptanceWhat the Clay prize rules require before a $1 million award○ open
Sources: Clay Mathematics Institute (Riemann Hypothesis page; Millennium Prize rules); Levent Alpöge, remark.pdf (November 20, 2026); Scientific American, November 20, 2026 (independent compile; comparator run awaited); Kevin Buzzard, Xena Project, November 12, 2026 (the checks). Status as reported in these sources by November 21, 2026.
Graphic: Powered by Thorsten Meyer AI
Why the Claim Matters for Number Theory
The Riemann Hypothesis is one of mathematics’ most prominent unsolved problems and one of the Clay Mathematics Institute’s Millennium Prize Problems, each carrying a $1 million prize for an accepted solution. It concerns the locations of certain zeros of the zeta function, a structure closely connected to how prime numbers are distributed.
A proof would settle the hypothesis and support many results in number theory that are currently conditional on it. But a large code archive changes the checking task rather than eliminating it: specialists must establish that the formalization matches the intended mathematics and that its result is sound. The episode also puts a sharp question before researchers: whether AI can contribute to difficult mathematics at a level that can be independently verified, rather than merely generate plausible-looking arguments.
Products Worth Considering

Brother P-touch EDGE PT-E310BT Industrial Label Printer

Vari ComfortEdge 48×30 Inch Adjustable Electric Standing Desk

Apple Mac mini with M2 chip, 16GB unified memory, 256GB SSD, Silver
As an affiliate, we earn on qualifying purchases.
From Riemann’s 1859 Problem to Lean
Bernhard Riemann introduced the conjecture in 1859. In the Clay Institute’s formulation, it says that every “non-obvious” zero of the Riemann zeta function has real part one-half. The conjecture matters because the zeta function encodes information about prime numbers, whose distribution appears irregular but can be described statistically.
Computers have checked the conjecture for the first 10 trillion solutions, according to the Clay Institute, but checking a finite number of cases cannot establish a claim about infinitely many. A proof must show that the statement holds generally. Lean can help verify formal steps mechanically, but experts still need to examine what theorem those steps encode and how the formal argument relates to conventional mathematics.
Alpöge’s account is the available source for how the work was divided. He says he directed the strategy and that Claude handled much of the technical development, writing that “Claude far outpaced me near the finish line.” He also says he has “little mastery” of one of the main areas used. Those descriptions have not yet been independently corroborated.
“I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis.”
— Levent Alpöge, in his November 20 note
The Proof’s Status and Open Checks
The central uncertainty is whether the Lean archive proves the Riemann Hypothesis as stated. Independent compilation indicates that the code can be checked under the reported axioms, but it does not by itself establish that the formalized theorem is the intended one. The comparator run described by Kevin Buzzard was still outstanding in the source account.
There is also no conventional proof available for mathematicians to assess in readable form. The note’s sketch is brief, and the claimed 250,000-page version was not posted. Terence Tao characterized the result as a Lean-only proof, while number theorist Andrew Granville called the released note “garbled,” according to the source account. The account says Anthropic had not announced the claim, had no related item on its news page on November 21, and had not immediately responded to Nature’s request for comment.
No journal has published the result, and no mathematical body has accepted it. The source material reports that a rival claim by a Stanford mathematician appeared the same day, but gives too little information to assess its substance or relationship to Alpöge’s claim. The identity of the Claude model and the precise contribution of the AI also remain unreported beyond Alpöge’s description.
Independent Verification Comes First
The next immediate milestone is the pending comparator check and further independent inspection of the Lean archive. Mathematicians will need to test both the formal code and its correspondence to the claimed theorems. The source account does not give a schedule for those checks or say who will conduct them.
Even if technical checks succeed, acceptance would take longer. The Clay Mathematics Institute’s rules require publication in a refereed journal, at least two years of scrutiny after publication, and general acceptance by the mathematical community before a prize is awarded. That process has not started on the information available: there is no journal publication or community acceptance reported. For now, the development is a notable claim with early verification activity, not a settled solution.
Key Questions
Has the Riemann Hypothesis been proved?
No accepted proof has been reported. Alpöge has made a claim, but mathematicians have not accepted it and independent checking remains incomplete.
What did Claude reportedly contribute?
Alpöge says he directed the strategy and Claude carried out much of the technical work, including a key step involving a chain of theta lifts. That division of work is Alpöge’s account and has not been independently verified.
What does the Lean code check?
Lean checks formal mathematical steps against stated rules and assumptions. A successful code check is not, on its own, proof that the formalized statement is the Riemann Hypothesis; that correspondence also needs review.
Has the Clay Institute prize been awarded?
No. The institute’s reported rules require publication in a refereed journal, at least two years of scrutiny and general acceptance. Those conditions have not been met for this claim.
Source: ThorstenMeyerAI.com
