Deicyde/riemann-hypothesis
Learning ResourcesA byte-exact Lean submission and Comparator setup for independently checking a claimed proof of the Riemann Hypothesis.
What this is
This repository packages a Lean formalization that claims proofs of the Riemann Hypothesis and GRH for Dirichlet L-functions, announced on 2026-11-20 in the remark PDF. It mirrors the submission unchanged and adds the files Comparator needs: a challenge file that states five theorems importing only Mathlib, a Comparator config, an export target list, and verification notes.
The five declarations cover Mathlib's RiemannHypothesis, an expanded restatement of it, non-vanishing of L(χ, s) for Re s > 1/2, zeros of L-functions in the critical strip lying on Re s = 1/2, and the primitive-character variant. This is infrastructure for verification, not a new mathematics write-up, and it does not claim that a run has passed.
Checking the claim
The process uses two Lean versions: the submission is built with Lean v4.33.1, while lake comparator requires Lean 4.35 or later and is run with v4.36.0. The exporter is lean4export tag v4.33.0 built against Lean v4.33.1. Full commands appear in the README's checking section; the outline is:
- In a clean clone at
$CLEAN, runlake exe cache get,lake build Challenge, then export the five challenge declarations withlean4export. - In a second, network-isolated clone, run
lake exe cache getbefore entering the sandbox, thenlake build Maininside it. - Export the solution from the sandbox to
solution.export. - From an empty directory with the config from the clean clone, run Comparator with
lake comparator, pointing at both exports.
The README warns to export the challenge first, before compiling the submission, because building runs the submission's code. The CI workflow does step 1 on every push and publishes the resulting challenge export as an artifact.
Requirements and limits
Expect hours of compute on a large machine, about 100 GB of disk, and substantial memory for the judging step. The verification notes record the source, the Lean version and its known kernel fixes, the tools and sandboxes used, a test of these steps, and what a passing run does and does not establish. A pass checks the exported proof against the challenge under Comparator's rules; it is not an independent mathematical peer review.
The submission is mirrored unmodified and was published without a license; rights remain with its author, and the repository's own README frames the mirror as for verification only.
HotRepo's take
The project appeared on HotRepo's new-project list on 2026-11-20, recording 102 stars and 19 forks during a roughly four-hour observation window that began the same day. That is a short window measured against a repository whose stated checking workflow takes hours on a large machine, so the star count reflects early attention around the announcement rather than a demonstrated verification result. The most useful material for readers is the reproducible Comparator path and the notes on kernel versions and sandboxes; whether the claimed proof holds depends on completing that run.
Written by AI from the project's README, updated . The project's own docs are the source of truth.
You might also like
- CocoSgt/riemann-hypothesis-leanA mirror of a claimed Lean 4 formalization of the Riemann and Generalized Riemann Hypotheses, for readers who want to build and check it.
- tristanbuckmaster/navier-stokesA paper and Lean 4 formalization of finite-time blow-up for the Navier–Stokes equations, aimed at mathematicians and proof engineers.