Anthropic mathematician says Claude helped him prove the Riemann hypothesis

Levent Alpöge's claim has not been independently verified, and Anthropic has not commented on the announcement or on the proof he posted on his own website.

· 2 min read

a blackboard with a lot of writing on it
Photo: Thomas T / Unsplash

Levent Alpöge, a mathematician at Anthropic, says he has proved the Riemann hypothesis with the help of the company's Claude model, The Verge reported. The problem is among the hardest in mathematics and has resisted leading researchers for decades. The claim has not been independently verified.

According to The Verge, Anthropic has not commented on the announcement and has not involved itself in it. The outlet said it contacted the company but did not immediately receive a reply. Alpöge wrote on his website that he messaged Anthropic CEO Dario Amodei about the result late on a Sunday night. He also suggests he waited to publish until after the company's IPO on Thursday.

In his statement, Alpöge says he believes he and Claude found a proof of the Riemann hypothesis. He adds that the method extends to the Generalized Riemann Hypothesis and, with more time, probably to the Extended Riemann Hypothesis. He also says the arguments have been formalized without assumptions. The solution is presented in Lean and has so far appeared only on his website.

The Riemann hypothesis was first proposed in 1859 and concerns how prime numbers are distributed, including the role played by the zeroes of a certain function. It is one of seven Millennium Prize problems. The Clay Mathematics Institute, which administers the prize, offers $1 million for each solved problem.

The institute has not publicly accepted Alpöge's work. Under its rules, a proposed solution must be published and withstand at least two years of scrutiny, during which it must gain general acceptance among mathematicians worldwide. In the roughly 25 years since the prizes were created, only the Poincaré conjecture has been officially solved.

The Verge noted that Anthropic said in August it was working on the Riemann hypothesis with Claude. At that point the company also said it did not expect the techniques Claude used to lead to a proof.

The claim arrives after a similar one from OpenAI. In September, that company said it had cracked the Navier-Stokes problem, which concerns fluid flow. According to The Verge, it spent millions of dollars and deployed 10,000 agents on the effort. The result is widely seen as credible but has not been formally recognized.

The Navier-Stokes work has also drawn allegations of questionable research practices, including claims that it drew on unpublished work shared by mathematicians with OpenAI's tools. OpenAI has strongly denied any wrongdoing. Alpöge had been working on Navier-Stokes with New York University professor Tristan Buckmaster. Buckmaster claimed that OpenAI's solution followed the route the two had been pursuing, using Codex and Claude.

Alpöge also said in July that he had solved the Jacobean conjecture with Claude. The report says the growing number of AI-assisted results has left mathematicians struggling to absorb the material and to work out their place in a rapidly changing field.

Related