AI-assisted and human-written Riemann Hypothesis proofs emerge in the same week
Anthropic mathematician Levent Alpöge says he found a Lean-checkable proof with help from Claude, while a Stanford postdoc posted a separate proof that Ravi Vakil says involved no AI.
· 2 min read
- Anthropic mathematician Levent Alpöge says he has a Lean-compilable solution to the Riemann Hypothesis, found with help from Claude.
- Stanford postdoc Alessandro Masullo separately posted a proof of the General Riemann Hypothesis that Ravi Vakil says involved no AI.
- Mathematicians still need time to understand both results, and Alpöge has committed to working on a human-readable explanation.
Two separate claimed proofs of the Riemann Hypothesis surfaced in the same week, one produced with the help of an AI model and one written entirely by a human, TechCrunch reported on Nov. 20, 2026 (UTC). The hypothesis dates to 1859 and is one of the Millennium Prize problems in mathematics.
The first claim comes from Levent Alpöge, a mathematician who works at Anthropic. The same week, his employer became the first publicly traded frontier AI lab. In an essay, Alpöge said he has a solution to the Riemann Hypothesis that compiles in Lean, and that the timing of the announcement was awkward. He wrote that he messaged Anthropic CEO Dario Amodei last Sunday night to ask how to make the finding public.
The hypothesis suggests that a hidden structure lies behind the way prime numbers appear. Lean is an open-source language used to check mathematical proofs, so a result that compiles in it can be verified mechanically. TechCrunch said most mathematicians seem confident that the formalized result will hold, although it will take time before the community fully understands the argument.
Experts have been working through the material in a chat on Zulip. Chris Birbeck, a number theorist at the University of East Anglia, wrote that the cohomology bound appears to be proved through an automorphy-lifting argument. He listed a long series of advanced tools the proof draws on, including Galois representations, base change and results by Deligne and Ribet.
Alpöge described his essay as a stream-of-consciousness account of assembling the pieces of the argument with Claude, after a great deal of confusion. He said he felt somewhat guilty about releasing the work without an explanation people can readily follow. He added that he has been fixated on the problem since childhood and felt he could not live with himself if he held it back. He has committed to helping produce a human-understandable explanation.
Alpöge and NYU mathematician Tristan Buckmaster also released a proof connected to the Navier-Stokes equations. According to TechCrunch, those problems helped start a race this summer between frontier AI labs and mathematicians to settle open questions.
The second claim came from Ravi Vakil, a Stanford professor and president of the American Mathematics Association. He posted a proof of the General Riemann Hypothesis written by Stanford postdoc Alessandro Masullo. Vakil noted that it is "all human, with no AI." The plan had been to bring Masullo to Stanford for a year to polish the writing and share it with enough people before announcing it, but the current draft was released now instead.
Masullo said the proof comes from twelve years of work. In his introduction he addressed mathematicians who feel discouraged by the pace and tone of the current climate, writing: "we are still here." Neither result has yet been confirmed by the wider mathematical community.