
How Lean Is Checking a Claimed Riemann Hypothesis Proof
A formal proof and several independent checks have been reported, but some verification runs were still underway at the latest update.
The short version
- Lean and Mathlib community representatives said they had certified the statement and proof of the Riemann Hypothesis submitted by Levent Alpöge. 1
- Lean is a proof assistant: software that checks whether a formal mathematical proof follows from specified rules and definitions. 3
- A verification page reported that the submitted source compiled and that several independent tools accepted the exported proof and compiled modules. 2
- As of the page’s November 21, 2026 update, two further checks were still in progress. 2
What does Lean verification of the Riemann Hypothesis mean?
Lean is both a proof assistant and a programming language. In this setting, a mathematician expresses a theorem and its proof in a precise form that software can check. The checking is not a judgment that the argument sounds convincing. It tests whether the formal steps follow from the definitions and foundational rules the system uses. 3
The community statement says Levent Alpöge released a Lean proof as a library on November 20, 2026. The statement says the Lean and Mathlib communities certified that both the theorem statement and the proof were correct. 1
What exactly was checked?
The verification page describes five headline theorems in the submission’s Main.lean file. One is the Riemann Hypothesis using Mathlib’s definition; another gives the same claim in the submission’s own wording. The page says Lean accepts those two statements as equal by definition. 2
The page also describes results about zeros of Dirichlet L-functions. These are mathematical functions related to the zeta function and number theory. The available source excerpt gives only part of the details for those theorems, so it does not support a fuller account of all five statements. 2
How do the independent checks help?
A proof assistant’s kernel is the small core that checks whether a formal proof is valid. Replaying a proof through a kernel is one way to check that the submitted files really produce a proof accepted by Lean. The MIT verification page reports that every module had been replayed through the Lean kernel. 2
That page also says two independently written kernels, nanoda and con-leche, accepted the full exported proof, and that lean4lean accepted every compiled module. A separate comparison mode matched all five theorems against a challenge file. These are different software checks of the formal artifact, not separate mathematical proofs written by different mathematicians. 2
Does this settle the Riemann Hypothesis?
The community statement says the proof and statement are correct, and the verification report describes substantial checking. 12 In that sense, the claim is not merely that someone has posted an informal argument: the proof has been expressed in Lean and accepted by multiple reported checking systems.
This does not mean every reader can easily inspect the argument. The community statement describes a library of 17 million lines of code, and says mathematical activity is still needed to explain the proof’s new ideas. 1 Formal checking can establish that a coded argument follows from its formal foundations; understanding why the argument works is a separate task.
What verification work was still unfinished?
Andrew Sutherland’s MIT page says its results were current as of November 21, 2026, at 12:55 UTC, and that the page would be updated as remaining checks finished. At that time, the lake comparator run and the Lean-kernel replay of the export were still in progress. 2
The Lean and Mathlib statement is more conclusive, saying the proof and statement had been certified. 1 The two sources describe different verification reports and snapshots. The available material does not say whether the two outstanding runs later finished, or provide their final results.
What we don't know yet
- Whether the two verification runs listed as in progress on November 21 later completed, and what they found. 2
- How mathematicians will explain the proof’s new ideas in an accessible form. The community statement says further work is expected on this. 1
- The provided sources do not give a detailed account of the proof’s mathematical argument.
Share this explainer
A browser that helps you stay safe
Sureno is a Chromium browser with plain-language privacy settings and an assistant that only reads a page when you ask.
Questions people ask
Has Lean proved the Riemann Hypothesis?
What is a Lean proof assistant?
Lean is software used to write and check formal mathematical proofs, and it is also a functional programming language. 3 It checks whether proof steps follow the system’s rules and definitions.
What part of the Riemann Hypothesis was verified?
The verification page identifies a theorem stating the Riemann Hypothesis using Mathlib’s definition, as well as an equivalent version written in the submission’s own words. It also lists results concerning Dirichlet L-functions. 2
Did independent software check the Lean proof?
The MIT verification page reports that two independently written kernels accepted the exported proof, while lean4lean accepted every compiled module. It also reports that a comparison tool matched all five headline theorems to the challenge. 2
Are all checks of the Riemann proof finished?
The latest time-stamped verification page in the supplied sources, updated November 21, 2026, said the lake comparator run and Lean-kernel replay of the export were still in progress. The sources provided here do not report their later status. 2
Sources
- Lean verification of the Riemann Hypothesis — Lean community blog, 2026-11-21
- RH Lean proof: verification files — math.mit.edu
- Lean (proof assistant) — Wikipedia











