Two Claimed Solutions to the Riemann Hypothesis Emerge: One Built With Claude, One Without AI
Mathematician Levent Alpöge, who works at AI company Anthropic, announced a Lean-compilable solution to the Riemann hypothesis. Around the same time, a Stanford postdoc's proof written without any AI was published.

Mathematician Levent Alpöge, who works at the AI company Anthropic, announced this week that he has apparently solved the Riemann hypothesis, one of the most important problems in mathematics. Formulated in 1859, it is among the Millennium Prize problems and conjectures that a hidden structure underlies the distribution of prime numbers. If the result holds, it would be a major mathematical achievement.
A proof in Lean
In an essay, Alpöge says he found the argument together with Anthropic's model Claude, and that it is expressed in Lean. Lean is an open-source language that encodes mathematical ideas as code so proofs can be formally verified. Several mathematicians in a discussion on the Zulip platform, including MIT's Andrew Sutherland, confirmed that the solution compiles successfully. After an update to the original report, independent sources have also checked it.
Experts are still working through the substance. Chris Birkbeck, a number theorist at the University of East Anglia, wrote in the chat that the cohomology bound appears to be proved by an automorphy-lifting argument. Most participants seem confident that the formalized result will stand. Alpöge acknowledges releasing the proof without an explanation humans can readily follow, saying his lifelong obsession with the problem meant he could not wait. He has committed to helping explain the result to other mathematicians.
A second proof, without AI
At the same time, Ravi Vakil, a Stanford professor and president of the American Mathematics Association, posted a proof of the General Riemann hypothesis written by Stanford postdoc Alessandro Masullo. Vakil stressed that it is entirely human work with no AI. The proof is not expressed in Lean, so it is too early to say whether it is correct. Masullo says it is the result of 12 years of work.
Alpöge and NYU mathematician Tristan Buckmaster have also released a proof related to the Navier-Stokes equations.
