Anthropic mathematician says Claude helped him prove Riemann hypothesis
Levent Alpöge says the proof compiles in the Lean system, but outside mathematicians have not checked it and the Clay Mathematics Institute has not accepted it.

Levent Alpöge, a mathematician on Anthropic's technical staff, said on Friday that he and the company's Claude AI model had found a proof of the Riemann hypothesis, one of the Clay Mathematics Institute's seven Millennium Prize Problems.
"I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis," Alpöge wrote in a statement on his personal website. He said the proof was formalized in Lean, a proof-assistant system, and that the code compiled.
The claim has not been independently verified, AFP reported. The proof has not been through outside peer review. The Verge reported that the Clay institute has not publicly accepted the solution. The Washington Post reported that Alpöge also claims a proof of the broader Generalized Riemann Hypothesis.
Alpöge wrote that he contacted Anthropic chief executive Dario Amodei last Sunday to discuss how to announce the result. He said "the precise timing is horrible" because Anthropic had just become publicly traded.
Formulated in 1859, the Riemann hypothesis is widely regarded as one of the central unsolved problems in mathematics. A valid proof would qualify for the institute's $1 million prize, subject to its rules and review.
The claim follows earlier AI-generated mathematics that drew skepticism. In October, OpenAI reportedly released hundreds of AI-generated mathematical manuscripts, including work on a weaker "quasi-Riemann hypothesis". Critics called for human verification. Alpöge's announcement differs in that it comes from a named mathematician at an AI lab and comes with a formal Lean artifact.
Compiling in Lean does not by itself show that the formal statement matches the Riemann hypothesis or that the translation contains no errors. Independent experts will have to examine both the code and the argument. The reports do not give a date for any Clay review.
Why it matters
If the proof holds, it would settle a 167-year-old problem and mark the first Millennium Prize Problem solved with an AI model's help. If it fails scrutiny, it will add to doubts about AI-generated claims in advanced mathematics, and Anthropic will face questions over a public claim made days after its stock market debut.
What’s next
- Independent mathematicians' review of Alpöge's Lean code and written argument
- Any statement from the Clay Mathematics Institute on whether it will review the claim
- Any formal statement from Anthropic on the claim
- Scrutiny of the Generalized Riemann Hypothesis claim reported by The Washington Post
Sources
- Using AI, mathematician claims to have solved one of math's great mysteries washingtonpost.com
- Anthropic mathematician claims to have solved prestigious Millennium ... theverge.com
- AI claimed a solution to the 167-year-old Riemann hypothesis — and a human did, too | TechCrunch techcrunch.com
- Mathematician says he and AI solved Riemann hypothesis nature.com
- When will an AI lab solve another Millennium Prize Problem | Octagon octagonai.co
- Will Anthropic announce a Millennium Prize solution | Octagon octagonai.co
- Anthropic researcher says Claude solved math's greatest riddle afp.com
- Has the Riemann hypothesis been solved? Anthropic mathematician Levent Alpöge claims Claude cracked math's most famous problem clarin.com
- AI claimed a solution to the 167-year-old Riemann Hypothesis — and // Tech Beat techbeat.co
- Anthropic worker claims to have solved a Holy Grail of math problems with the help of AI ctpublic.org
Reported and written by Tremor’s AI newsroom from the public sources above. Check them before you act on anything here.
More in Science
All science- Yesterday
Mathematicians claim proof Navier-Stokes flows can blow up without force
The paper by an NYU professor and an Anthropic researcher has not been verified, and the Clay Mathematics Institute has not declared the Millennium Prize problem solved.
- Nov 15
SpaceX ties Doge-1 lunar launch timing to Starship V3 tests
The company's mission update gives no launch date, and the dogecoin-funded satellite was first promised for early 2022.
- Oct 27Oct 27
Review panel confirms AI-generated proof of Yang-Mills existence; mass gap open
The result covers only the existence half of the Clay Millennium Prize problem, so the full problem remains unsolved.