The people who run Lean, the software mathematicians use to check proofs by computer, say a proof of the Riemann Hypothesis is correct. “We hereby certify that both the statement and the proof of the Riemann Hypothesis are correct,” reads a joint statement by the Lean FRO and the Mathlib Initiative posted on Saturday.
The proof is by Levent Alpöge, a mathematician at Anthropic, who said on Friday he found it working with Claude. Riemann posed the problem in 1859, 167 years ago by our arithmetic.
We track what AI systems can do in mathematics and other tests on our AI model benchmarks page.
Business Pill: The Cost of Checking
An evergreen Business Engineer explainer on the cost of checking work. It teaches the general idea only and does not discuss the Riemann proof.
The key insight: The Lean FRO and the Mathlib Initiative certified on 21 Nov that Levent Alpoge’s 17-million-line Lean proof of the Riemann Hypothesis, posed 167 years ago, is correct; Alpoge, an Anthropic mathematician, says he found it working with Claude.
What the Lean Statement Says
Alpöge released the proof on 20 November as a Lean library of 17 million lines of code, the statement says. “We have full confidence that Alpöge’s proof of the Riemann Hypothesis is correct, and we congratulate him on this achievement.”
It lists four checks. The Riemann statement used is Mathlib’s own, “carefully vetted by experts in Lean and number theory”. Dozens of people compiled the code and got the same proof file.
Two independent tools confirmed the theorem proved is exactly Mathlib’s Riemann statement. Several proof checkers, including Lean’s official kernel and checkers written by third parties in other languages, accepted the proof.

What a Computer Check Means
In plain words, a computer program has checked every logical step, down to the basic rules of logic. It does not get tired or skip a line, which is the risk with a human referee reading a very long proof.
The proof uses only Lean’s three standard axioms, according to a verification page kept by Andrew Sutherland of MIT, who says he ran the checks with Claude Code. All 175,187 modules compiled and passed the kernel replay, his page says.
The Riemann Hypothesis says every non-trivial zero of the zeta function has real part one half. It governs how prime numbers are spread out.
What Alpöge Says Claude Did
“I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis,” Alpöge wrote in a short note. He says the method also resolves the Generalized Riemann Hypothesis.
He credits Claude with a key step: “this time Claude far outpaced me near the finish line.” This account of Claude’s role is his own; the Lean statement certifies the proof, not who found which step.
The Structural Read
As we read it, the news is the checking as much as the proof: a computer verified every step within about a day of release, which a human referee could not do for 17 million lines.
As we read it, the open questions are human ones: a readable explanation of the new ideas, and the Clay process, which by its rules cannot conclude for at least two years after publication.
Lean FRO and Mathlib Initiative, 21 Nov 2026
“We hereby certify that both the statement and the proof of the Riemann Hypothesis are correct.”
Three Implications
FOR MATHEMATICIANS the proof is machine-checked but not yet explained in a readable paper.
FOR AI LABS Alpoge credits Claude with a key step; the certification covers the proof itself.
FOR READERS the Clay prize requires publication, two years and general acceptance.
What Is Not Established
There is no human-readable paper yet. Alpöge calls the argument “incredibly long at the moment”. The Lean teams say they look forward to work “that will help explain the new ideas in this proof”.
The broader claim is narrower than it sounds. Sutherland’s page says the generalized result checked covers Dirichlet L-functions only, not other families.
The $1 million Clay Millennium Prize is not decided. Clay’s rules require publication in a qualifying outlet, at least two years since publication, and general acceptance by mathematicians.
The Bottom Line
The Lean FRO and the Mathlib Initiative certify that a 17-million-line computer proof of the 167-year-old Riemann Hypothesis is correct. Its author credits Claude; the human write-up and the Clay prize process are still ahead.
97,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. The certification and its four checks come from the Lean statement, read in full. Check counts and the axiom list come from Andrew Sutherland’s verification page; some checks there were still running when it was last updated. Claude’s role is as Alpöge describes it in his note. The 167 years and the other year counts are our arithmetic.
Sources: Lean FRO and Mathlib Initiative, ‘Lean verification of the Riemann Hypothesis’, Lean community blog, 21 Nov 2026 · Andrew Sutherland (MIT), verification files for the Lean proof of the Riemann Hypothesis, updated 21 Nov 2026 · Levent Alpoge, remark announcing the proof, 20 Nov 2026 · Clay Mathematics Institute, Millennium Prize rules · Andrew Curran on X, 21 Nov 2026 · FourWeekMBA on the Riemann claim, 20 Nov 2026









