Route · Lean proof of the Riemann hypothesis
How the Lean proof of the Riemann hypothesis is structured: the route to the main theorem
The Lean proof of the Riemann hypothesis is a library of 175,187 Lean files (175,027 mapped), grouped on this map into 2,647 clusters. A cluster's depth is the number of import steps from the main theorem's cluster to it.
Proof status as of 22 Nov 2026: on 21 Nov 2026 a joint statement by the Lean and Mathlib communities, endorsed by the Lean FRO and the Mathlib Initiative, certified the Lean statement and proof. More independent checks: math.mit.edu/~drew/grh. Names and descriptions on this map are written by Claude from samples of the code and may be wrong.
Clusters by depth
Rings around the main theorem, from the centre out:
The first two steps
- ★ Main theorem: Main theorem statements, plus the explicit formula for Dirichlet L-functions
- Primes in progressions, Siegel-Walfisz and exceptional zero bounds depth 1
- Zero-free region for zeta and Dirichlet L-functions depth 1
- Mertens-type estimates and zero sums for Dirichlet L-functions depth 1
- Cuspidal spectrum, trace class and published representation theory depth 2
- Hecke characters, conductors and functional equations over CM hosts depth 2
- Weight-indexed polynomial spaces, lattices and branching over number fields depth 2
- Norm sums, size bounds and the archimedean height bound depth 2
- Family bounds on shares, degrees and admissible constants depth 2
- Signatures of hermitian forms, Hasse norm theorem, slopes depth 2
- Host field data, class number lemmas and Galois descent for embeddings depth 2
- Test lattices, dual lattices and slope bounds depth 2
- Explicit parameter choices: scale, ranks and host conditions depth 2
- Chain parameters, weighted means and grid valuation bounds depth 2
- Division values, index change and divisor bounds depth 2
File by file
- 175,187 Lean files, all reached from
Main by imports - Median 13 imports from
Main; farthest 43 - Longest import chain: 203 files
Distances are shortest import chains; the cluster graph has loops, the file graph does not.
Open the map · start at the main theorem · every cluster as a list, by distance