Newzly.net

technology·California·San Francisco County

Anthropic mathematician says Claude helped prove the Riemann hypothesis

Levent Alpöge says the AI system helped him develop a proof of the prime-number problem over about six weeks, according to a note he posted Friday.

2 min readvia The Guardian Technology
Illustration of mathematical notes on prime numbers beside a computer, suggesting an AI-assisted proof claim
Illustration of mathematical notes on prime numbers beside a computer, suggesting an AI-assisted proof claim

In this piece

A mathematician at Anthropic says Claude helped him prove the Riemann hypothesis, a central unsolved problem about prime numbers. Levent Alpöge posted a note…

A mathematician at Anthropic says the company’s Claude system helped him prove the Riemann hypothesis, long regarded as one of the most important unsolved problems in mathematics. Levent Alpöge, a Turkish American mathematician, posted a note on his personal website on Friday describing the claim, The Guardian reported.

Alpöge wrote that Claude helped develop a solution over about six weeks. The note, linked from his site as a PDF, is the basis of the account. The Guardian’s report presents the work as his contention, not as a result already accepted by the wider mathematical community.

The Riemann hypothesis concerns the Riemann zeta function, introduced by Bernhard Riemann in the nineteenth century, and the location of its so-called non-trivial zeros. In broad terms, it predicts that those zeros lie on a single vertical line in the complex plane. A proof would tighten mathematicians’ understanding of how prime numbers are distributed among the integers.

Primes — numbers divisible only by one and themselves — sit at the center of number theory and appear in practical systems such as cryptography. The hypothesis does not list new primes, but it governs how sharply mathematicians can estimate how many primes fall below a given size. That link is why a claimed proof draws attention well beyond specialist journals.

Alpöge’s account, as described by The Guardian, says he used Claude as a working partner while pursuing the argument. The reporting available here does not reproduce a line-by-line proof, name outside referees, or quote Anthropic endorsing the result as settled. Readers have his published note and a news report of it, rather than a completed community verdict.

Anthropic, the San Francisco-based company that builds Claude, employs Alpöge as a mathematician, according to the Guardian account. The article does not say the firm commissioned the proof or that it has reviewed the manuscript. The claim is framed as his, posted on a personal site.

Extraordinary claims in pure mathematics are settled by scrutiny, not by the identity of the tools used along the way. A proof of this scale would normally be read for gaps, checked against known partial results, and, if it holds, published in a peer-reviewed venue. None of those later steps is described as finished in the material available here.

The six-week timeline he cites is short relative to the problem’s age, which dates to Riemann’s 1859 paper. Speed alone neither confirms nor refutes the work. It does mean specialists will look closely at which steps were human, which were machine-suggested, and whether every inference is fully written down.

The Guardian story, filed in its technology section in late November 2026, rests on Alpöge’s Friday note. It does not report independent verification, a formal comment from Anthropic, or responses from leading number theorists. Until such checks appear, the accurate description is a claimed proof assisted by Claude, not a settled theorem.

What happens next is outside review of the PDF he posted. If referees find a complete argument, the result would rank among the landmark achievements in mathematics and a notable use of a large language model in research. If they find a gap, the episode will still show how quickly AI tools are being brought into problems that have resisted mathematicians for generations.

Primary source: The Guardian Technology. Newzly rewrites and structures source reporting for clarity; verify critical details at the original.