Lean community confirms Levent Alpöge's proof of the Riemann hypothesis
A Stanford mathematician has published a machine-checked proof of the Riemann hypothesis, and dozens of independent verifiers have confirmed it is correct. The breakthrough arrived years before most forecasters expected.

Key takeaways
- The Lean and Mathlib community has formally confirmed Levent Alpöge's proof of the Riemann hypothesis.
- The proof was published on 20 November 2026 as 17 million lines of Lean code and checked independently by dozens of verifiers.
- The verification software is designed so it provably cannot accept a flawed proof.
- Forecasters had expected a solution around May 2035, and Terence Tao predicted one around 2050.
In this article
A machine-checked proof wins official approval
The Lean and Mathlib community has given its formal approval to a proof of the Riemann hypothesis by Stanford mathematician Levent Alpöge, according to The Decoder. In a statement, the community said it has full confidence that the proof is correct and congratulated Alpöge on the achievement.
Alpöge published the work on 20 November 2026. Rather than a traditional paper, the proof takes the form of 17 million lines of code written in Lean, a software tool that checks mathematical arguments for errors line by line.
How the verification worked

Dozens of people ran the code independently and each obtained the same result. The verification programs are built so that they provably cannot accept a flawed proof, which removes the usual risk of human oversight in peer review.
MIT mathematician Andrew Sutherland lists the independent checks on his website, providing a public record of the verification effort. This kind of formal verification, where a computer confirms every logical step, has become an increasingly common way to settle contested mathematical results.
Why the Riemann hypothesis matters
The Riemann hypothesis had long ranked among the most famous unsolved problems in mathematics. It makes a prediction about how prime numbers are distributed, and a proof would firm up results across number theory that currently rest on the assumption it is true.
Because the proof was checked entirely by machine, the mathematical community can be certain of its validity in a way that is rarely possible for work checked only by human referees.
A result that arrived far ahead of schedule
The solution came much earlier than specialists anticipated. Forecasters on the prediction platform Metaculus had put the expected date at May 2035, as Ethan Mollick notes.
The actual outcome beat even that generous timeline by more than two decades.
Sources
Spotted a mistake? See our corrections policy. Read how News by OnDevice works in our editorial policy.



