KEY POINTS
- Anthropic mathematician Levent Alpöge claims he proved the Riemann hypothesis with help from Claude.
- The proof, written in Lean and posted on his website, has not been independently verified, and Anthropic has not commented.
- Under Clay Mathematics Institute rules, a solution must be published and withstand at least two years of scrutiny before it can be considered for the prize.
Levent Alpöge, a mathematician at Anthropic, says he has proved the Riemann hypothesis with the help of the company's Claude model, according to The Verge. The problem has resisted some of the world's best mathematicians for decades. His claim has not yet been verified by anyone outside the project.
Alpöge laid out his claim in a document on his personal website. He writes that he believes he found a proof working with Claude, and that the method extends to the Generalized Riemann Hypothesis. He also expects, given more time, that it will cover the Extended Riemann Hypothesis. He adds that he and Claude formalized the arguments without assumptions, and the solution is presented in Lean.
Anthropic has not commented on the announcement or taken part in it, The Verge reported. Alpöge says he texted CEO Dario Amodei about the result "late Sunday night". He also suggests he waited to publish until after the company's IPO on Thursday. The Verge said it contacted Anthropic and had not received an immediate reply.
The Riemann hypothesis was first proposed in 1859 and concerns how prime numbers are distributed, and the role zeroes of a certain function play in that pattern. It is one of seven Millennium Prize problems. Anthropic said in August that it was working on the problem with Claude, but added that it did not expect the techniques Claude used to lead to a proof.
Even if the proof holds up, a prize is far off. The Clay Mathematics Institute, which administers the awards and offers $1 million per problem, requires that a solution be published and survive at least two years of scrutiny. During that time it must gain general acceptance in the global mathematics community. So far only one Millennium problem, the Poincaré conjecture, has been officially solved.
The claim arrives amid a run of AI-linked results in mathematics. In September, OpenAI said it had cracked the Navier-Stokes problem on fluid flow. The company says it acted after hearing that rival researchers were close to a solution. The Verge said OpenAI spent millions of dollars and deployed 10,000 agents. Its result is widely seen as credible but has not been formally recognized.
That effort has also drawn controversy. Allegations of questionable research practices include claims that OpenAI used unpublished work from mathematicians who had shared it with its AI tools. OpenAI has strongly denied any wrongdoing. The report adds that Alpöge had been working on Navier-Stokes with New York University professor Tristan Buckmaster. Buckmaster has claimed OpenAI's solution followed the route the two had been pursuing.
Alpöge also said in July that he had solved the Jacobean conjecture using Claude. The Verge reports that the steady stream of results has left mathematicians struggling to absorb the new material. They are also trying to work out where they fit in a field that is changing quickly.