Claude Helped With The Riemann Hypothesis, Says Anthropic Mathematician
AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: Claude Helped With The Riemann Hypothesis, Says Anthropic Mathematician on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get the latest gadgets delivered free with Prime

  • Fast, free delivery on millions of items
  • Prime Video, Amazon Music and more included
  • Member-only deals all year
Start your free Prime trial Free trial for eligible customers · Cancel anytime
As an affiliate, we earn on qualifying purchases.

TL;DR

Anthropic mathematician Levent Alpöge says he worked with Claude on a proof of the Riemann hypothesis and a broader related conjecture, releasing a Lean code archive rather than a readable proof. Some initial checks have been reported, but a key comparison test was still pending and the claim has not been accepted by mathematicians.

Levent Alpöge, a mathematician who works at Anthropic, said on November 20 that he believes he and the company’s Claude AI model have proved the Riemann hypothesis, one of mathematics’ best-known unsolved problems. He released a brief note and a large archive of Lean code, but the proof has not been accepted, and mathematicians are still checking whether the code establishes the claimed result.

Alpöge’s note, posted on his personal website, says the approach also extends to the Generalized Riemann Hypothesis, a broader family of related claims. The note is not a complete, conventional proof: it sketches some ideas, provides a timeline, and links to a Lean archive divided into 14 parts. Lean is software that checks whether a formalized mathematical argument follows from its stated foundations.

According to the source account, computer scientist Scott Aaronson put the archive at 17.4 million lines of Lean code. Alpöge said a human-readable version would exceed 250,000 pages and chose not to publish it. He characterized himself as having guided the strategy and Claude as having handled much of the technical work, writing that “Claude far outpaced me near the finish line.” The account does not identify which Claude model was used.

Initial checks offer limited information, not a verdict on the mathematics. MIT mathematician Andrew Sutherland told Scientific American that the code compiled independently “with only the three standard axioms.” Kevin Buzzard said a “comparator run” was still awaited; that test is intended to check that the formalization proves the intended statement. Terence Tao described the result as an “(enormous) Lean-only proof.” James Maynard told Nature, “If this fully checks out, it is clearly a historic milestone in mathematics.” Andrew Granville, by contrast, called the released note “garbled.”

At a glance
reportWhen: Claim posted November 20, 2026; scrutin…
The developmentLevent Alpöge posted a note saying he believes he and Claude proved the Riemann hypothesis, alongside a large Lean formalization that mathematicians are beginning to check.

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

0 1/2 1 real part

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

  1. 1Claim postedA 4-page note plus Lean code, November 20, 2026✓ done
  2. 2Code compiles elsewhereIndependent compile reported by an MIT mathematician✓ reported
  3. 3Comparator runRight statement? Only standard axioms? Several Lean kernels?… pendingWe are here · November 21, 2026
  4. 4Humans understand itA readable proof that experts can follow○ open
  5. 5Journal publicationA refereed mathematics journal of worldwide repute○ open
  6. 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 a Proof Would Reshape Number Theory

The Riemann hypothesis concerns the locations of special zeros of the zeta function and their relationship to the distribution of prime numbers. A proof would settle a problem that has remained open since Bernhard Riemann proposed it in 1859. Many results in number theory are currently established only on the assumption that the hypothesis is true, so confirmation could turn conditional results into unconditional ones and support further work across the field.

The claim is also a test of AI-assisted mathematics. A Lean formalization can let software verify each encoded logical step, but a successful compilation alone does not establish that the formalized statement matches the claimed theorem or that the argument has been represented correctly. Independent experts would need to determine whether the archive proves the result Alpöge says it proves.

The potential financial prize is not yet in play. The Clay Mathematics Institute lists the Riemann hypothesis among its Millennium Prize Problems, each carrying a $1 million prize. Its rules require publication in a refereed journal, at least two years of scrutiny, and general acceptance by the mathematical community. None of those conditions has been met by the material described so far.

Products Worth Considering

LG 27GS60QC-B Ultragear 27-inch Curved Gaming Monitor (QHD, 180Hz, 1ms, 1000R)

LG 27GS60QC-B Ultragear 27-inch Curved Gaming Monitor (QHD, 180Hz, 1ms, 1000R)

As an affiliate, we earn on qualifying purchases.

From Riemann’s Conjecture to Lean

The hypothesis says that all the “non-obvious” zeros of the Riemann zeta function have real part one-half. In simplified terms, it predicts a deep regularity in how prime numbers are distributed. The Clay Mathematics Institute says the claim has been checked for the first 10 trillion solutions, but checking a finite number of cases cannot prove a statement about infinitely many cases.

Alpöge’s publication presents the claimed proof principally as machine-checkable code rather than a paper that mathematicians can read and assess directly. The source account says he also produced a human-readable document of more than 250,000 pages but did not post it, describing such a version as a “travesty.” The note’s short sketch does not provide a full account of the argument.

The claim has drawn attention because its author works at Anthropic and says Claude helped. The source account reported that Anthropic had not announced the claim on its news page as of November 21, and Nature reported that the company had not immediately responded to its request for comment. The claim remains Alpöge’s, not an announcement or endorsement by Anthropic.

“I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis.”

— Levent Alpöge, in his note

The Formalization Still Needs Scrutiny

The central open question is whether the Lean archive proves the Riemann hypothesis itself and, separately, the claimed generalized version. The reported independent compilation is an early check; the comparator run mentioned by Buzzard had not yet been completed in the source account. Its result, and any full expert assessment of the code, remain unknown.

Alpöge’s account is currently the available description of how the work was divided between him and Claude. The source material does not establish which model was used, provide an independently reviewed explanation of the proof, or report formal acceptance by a journal or mathematical organization. It also does not establish whether other mathematicians can reconstruct and understand the argument from the code.

Anthropic had not publicly commented in the account available on November 21, and its role beyond employing Alpöge is unclear. There is also no indication that the Clay Institute has evaluated the submission or that its prize process has begun.

Independent Checks Will Decide the Claim

The immediate next step is for mathematicians to complete the reported comparator run and examine what the Lean archive formally proves. Further scrutiny will need to establish that the code corresponds to the intended theorem and that its foundations and dependencies are acceptable to experts. Any errors, limitations, or successful verification would change how the claim is assessed.

For the result to gain formal standing, Alpöge or others would also need to make the work available for sustained mathematical review and meet the publication and community-acceptance requirements described by the Clay Institute. Until then, the accurate description is that an Anthropic mathematician has claimed a Claude-assisted proof and released code for checking—not that the Riemann hypothesis has been solved.

Key Questions

Has the Riemann hypothesis been proved?

Not on the information reported so far. Alpöge says he believes he and Claude proved it, but mathematicians have not accepted the claim and checks are ongoing.

What did Alpöge release?

He posted a short note with a sketch and a link to a Lean archive in 14 parts. The source account describes the archive as 17.4 million lines; it is not a conventional readable proof paper.

What does Lean checking show?

Lean can check whether a formalized argument follows logically within its stated framework. A compilation does not, on its own, establish that the formalization proves the intended theorem; experts still need to check that connection.

Did Claude do the mathematical work?

Alpöge says he guided the strategy and Claude carried out much of the technical work. That account has not been independently established, and the source material does not identify the Claude model used.

Does the claim qualify for the $1 million Clay prize?

Not at this stage. The Clay Institute’s rules require publication in a refereed journal, at least two years of scrutiny, and general acceptance by mathematicians.

Source: ThorstenMeyerAI.com

You May Also Like

AI Tracker Boosts Wide-Area Surveillance with 42% Fewer Identity Switches

AIThis post was created with the assistance of artificial intelligence (AI).The published…

Jack Clark Says It Out Loud — Reading the Co-Founder’s 60%/2028 Estimate on Automated AI R&D

Anthropic’s co-founder Jack Clark publicly estimates over 60% probability that autonomous AI R&D occurs by 2028, signaling a major policy stance.

How and when to watch May’s blue moon

Learn the confirmed time and viewing tips for May’s blue moon, the second full moon in a month, visible across various regions this weekend.

OlmoEarth Embeddings: Precision Data Exports For AI Applications

OlmoEarth Studio now supports on-demand export of satellite data embeddings, enabling advanced AI applications like similarity search and land classification.