Levent Alpöge, a mathematician on Anthropic's staff, announced on Friday that he and Claude have found a proof of the Riemann Hypothesis, the 1859 conjecture on how the prime numbers are spread out and one of the Clay Mathematics Institute's seven $1 million Millennium Prize problems.

"I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis," Alpöge wrote in a four-page note on his website. He said the method "easily extends much further," settling the Generalized Riemann Hypothesis as well, and that he expects the Extended Riemann Hypothesis to follow "with more time."
Instead of a paper, Alpöge released the proof written in Lean, a language in which a computer checks every step of an argument, split into 14 downloadable archive files. "Claude and I have also formalized the arguments without assumptions," he wrote, adding that the argument "is incredibly long at the moment."
He said he chose to publish only the Lean code because "it being obviously inappropriate to attempt to post some massive unreadable travesty of 'slop', I decided to only present Lean." In a footnote he admitted he had tried anyway: "I of course amused myself, and hit 250,000+ pages."
As of Friday, no outside mathematician had vouched for the result. - and number theorists still have to confirm that the theorem at the end of the code really is Riemann's hypothesis, as Scientific American notes.
Alpöge, a number theorist, wrote that he had been planning for this problem since his student days, and gave Claude the decisive turns. "Claude and I have had a great time together with such breadth in recent months, but this time Claude far outpaced me near the finish line," he wrote. "No, Claude flew, and danced, and sparkled near that finish line."

He said he steered the work toward the right setting and that Claude then supplied the step he expects to make mathematicians "fall out of their chairs": a chain of "theta lifts" about as long as the square root of the size of the prime sum being bounded, each lift standing in for the "multiplicity" that the classic argument over function fields relies on and that is hard to find over the whole numbers. A timeline he says Claude compiled runs from Oct. 6, the "Stepanov determinant," to Nov. 16, with a single entry: "GRH."
Why does it matter? Take the first trillion whole numbers. A standard formula predicts about 37.6 billion primes and misses by just 38,262. If the proof holds, mathematicians would have a guarantee on how far that estimate can stray, even at numbers far too large to check by counting. Results that depend only on Riemann being right would become theorems. For Anthropic, the bigger prize would be evidence that Claude contributed an original idea that helped crack a problem mathematicians have chased since 1859.
Alpöge's claim landed a day after Anthropic's first day of trading, with the stock sliding toward its $83 IPO price on Friday morning. He addressed the timing in his note - writing: "It is obvious to me that I should remark on the following." "The precise timing is horrible," adding that it all came together late Sunday night, when he messaged Anthropic CEO Dario Amodei, "Hypothetically, if I were to have ...", "eventually asking about my ability to announce."
"We went and figured it out and the fact is that there was a constraint to respect," he went on, adding that he was "clearly unilaterally responsible for feeling I just could not live if I waited even longer." "There is an obvious cynical view, to do with the timing, but it is false, and I am pleased that I am in circumstances where it is obviously false."
In August, Anthropic said an unreleased version of Claude had raised the share of the zeta function's zeros known to satisfy the hypothesis from 41.6% to 67.2%, in work that Alpöge and fellow Anthropic mathematician Ralph Furman validated. The company wrote at the time: "We don't expect that the techniques Claude used will lead to proving the Riemann hypothesis." In July, Alpöge posted a Claude-generated counterexample to the Jacobian conjecture.
Even if the proof holds, the money is years off: under Clay's rules, a solution must be published in a qualifying outlet, wait at least two years and win "general acceptance in the global mathematics community" before the institute will consider it.
Bettors aren't sold yet:
Alpöge says he will write the argument up properly and expects it to shrink fast. "My intuition insists it will be massively simplified, likely immediately," he wrote. "It is natural to feel that this is an end, in my case a loss of a principal guiding light, but what I saw during the course of progress so amazed me that I believe it is one dawn across science of something far greater."