Anthropic mathematician says Claude helped prove the Riemann hypothesis
Bloomberg reports that an Anthropic researcher believes he used Claude to solve the Riemann hypothesis. The proof is unverified, and no Anthropic statement has been reported.

An Anthropic researcher says he used the company’s Claude model to solve the Riemann hypothesis, a number theory problem that has resisted mathematicians for more than 150 years, Bloomberg reported. The claim is unverified, and the researcher himself says only that he believes he has a proof.
Bloomberg ties the news to Anthropic’s stock market debut, saying it came on the heels of the company’s record-breaking initial public offering. The outlet’s public summary goes no further, so the details below come from other coverage we read.
What the claimed proof consists of
Interesting Engineering reports that the researcher is mathematician Levent Alpöge. According to that report, the work extends to the Generalized Riemann Hypothesis, a broader version of the conjecture, and he credited Claude with making rapid progress in the final stages. Techmeme, summarizing Scientific American, says he posted the proof in the Lean programming language.
Lean is a proof assistant: it checks every step of an argument against formal rules. Interesting Engineering describes the release as a short written explanation plus a very large set of Lean verification files.
We could not open the researcher’s own page or the Lean archive, and none of the reports we read links to it. Everything here about the proof is secondhand.
Why compiling code is not yet a proven theorem
Code that compiles does not by itself show that the right statement was proved. Interesting Engineering notes that researchers must still confirm that the Lean theorem matches the conventional Riemann hypothesis, check the definitions and look for unsupported assumptions. It reports that an independent verification project confirmed the downloaded files match the published archive and that the main statements compile against Mathlib, Lean’s community math library. That project is not named, and no mathematician is quoted endorsing the result.
The prize question is also open. The Clay Mathematics Institute lists the Riemann hypothesis among its Millennium Prize Problems. According to Interesting Engineering, the rules require publication in an approved peer-reviewed outlet followed by a waiting period of scrutiny and broad acceptance, so the announcement does not qualify for immediate consideration, and the problem remains officially unsolved.
The Verge, the Financial Times and The Information also carried the story, but their texts were not available to us. We found no Anthropic statement in the coverage we could read, so the company’s position on the claim is unknown.
A claim landing in the middle of an IPO week
The timing matters for readers following Anthropic as a public company. Our earlier coverage reported on the pricing and Nasdaq debut of the IPO and on how the offering closed once chipmakers’ private stakes were counted. A genuine proof would be a landmark for AI in mathematics. An error, or a Lean statement that does not match the real conjecture, would be a very public setback. Nothing in the sources links the announcement to the share price.
For now the claim is only a claim. The measurable next steps are independent checks of the Lean statements, a written paper that mathematicians can read, and any response from Anthropic. We will update this report when any of them appears.
Sign Up for the Daily AI Pulse
One email a day. All the stories that matter.
By signing up, you agree to our Terms of Use and acknowledge the data practices in our Privacy Policy. You may unsubscribe at any time.
Alex CarterAI News & Big Tech CorrespondentAlex Carter writes for Aiholics, keeping readers updated on the fast-paced world of AI and Big Tech.



