
Independent Mathematicians Confirm AI Proof of Riemann Hypothesis
Mathematicians from the Lean and Mathlib communities say Levent Alpöge's Claude-assisted proof of the Riemann Hypothesis is correct, after independent compilation and a completed comparator run.
Mathematicians are now certain that the Claude-assisted proof of the Riemann Hypothesis is correct. The hypothesis, first posed by Bernhard Riemann in 1859, concerns the distribution of the prime numbers. On 21 November 2026 the Lean and Mathlib communities certified, in a joint statement endorsed by the Lean FRO and the Mathlib Initiative, that both the statement and the proof are right, one day after Levent Alpöge released the 17-million-line Lean library on 20 November. "We have full confidence that Alpöge's proof of the Riemann Hypothesis is correct, and we congratulate him on this achievement," the statement says. "We hereby certify that both the statement and the proof of the Riemann Hypothesis are correct."
The four criteria
The statement is Mathlib's own RiemannHypothesis, and experts in Lean and number theory have vetted it carefully. Alpöge's library depends directly only on Mathlib, compiles without errors or build failures, and dozens of people have reported reproducing the same raw proof export file. To rule out a mismatch, the community added a file that isolates the statement and imports only Mathlib; two independent tools, comparator and con-leche, compared it with the theorem proved in the library and both confirmed that it is exactly Mathlib's statement of the Riemann Hypothesis. The exported raw proof was then checked by several independent checkers, including Lean's official kernel, con-leche and con-ron, and third-party checkers implemented in different programming languages. con-leche and con-ron come with a formal proof, relative to set theory, that they cannot accept a false proof. All the checks ran on an inert export of the proof, out of reach of hypothetical malicious code in the Lean sources.
How the doubts were resolved
Alpöge first checked the work by writing the proof out in Lean; a Lean proof that compiles without errors is usually taken as confirmation that the logic and the mathematics are right at every step. Kevin Buzzard of Imperial College London asked in a blog post, however, whether the theorem stated in the Lean code is the actual Riemann Hypothesis. Independent researchers, including Andrew Sutherland of the Massachusetts Institute of Technology, have now compiled Alpöge's result separately without problems. "The Lean proof has been compiled and independently rechecked by the Lean kernel with only the three standard axioms," Sutherland said. "Honestly the chance that anything is fishy here is approximately zero." The comparator run, which usually takes a matter of hours, has now been completed.
What the certification does not settle
The joint statement says nothing about the rival human draft that Stanford published after Alpöge's first claim, which sitech.ge reported earlier. The authors also note that the new ideas in the proof still have to be explained: "We look forward to all the activity in the mathematical community that will help explain the new ideas in this proof."
Earlier coverage: Anthropic Claims AI Proof of Generalized Riemann Hypothesis, Stanford Releases Human Draft
Sources: Scientific American · Lean Community Blog · Techmeme
SiTech — AI-powered web development
We build fast, modern websites and bring AI into real business workflows. Have a project or a question? We'd love to help.