Key Notes
- Levent Alpöge says he and Claude have found a proof of the Riemann hypothesis, a claim that still requires independent verification.
- His announcement links to a Lean formalization, while a conventional mathematical explanation is still to come.
- The claim goes far beyond Claude’s earlier partial result and does not mean the $1 million Millennium Prize has been awarded.
Levent Alpöge, a mathematician at Anthropic, says he and Claude have discovered a proof of the Riemann hypothesis, a 167-year-old problem at the heart of number theory. If independently confirmed, the result would rank among the most consequential mathematical achievements of the modern era.
In a four-page statement posted on his website on November 20, Alpöge describes a completed formalization and links to a Lean archive divided into 14 parts. AIstify has not independently verified the formal proof. The Clay Mathematics Institute still lists the problem as unsolved.
“I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis,” he writes.
Why the Riemann Hypothesis Matters
The hypothesis concerns the distribution of prime numbers, the building blocks of whole-number multiplication. Mathematicians can estimate how many primes lie below a given number; the Riemann hypothesis would sharply constrain how far the actual count can deviate from that estimate.
Its precise statement concerns the Riemann zeta function: every nontrivial zero must have a real part equal to one-half. Clay explains that checking vast numbers of individual zeros, however convincingly, is different from proving the statement for all of them. The conjecture dates to Bernhard Riemann’s 1859 paper.
That distinction also separates a successful numerical experiment from the result Alpöge is claiming. His announcement concerns a general proof, rather than a larger collection of examples that fit the prediction.
Claude’s Role and the Verification Ahead
Alpöge credits Claude with moving beyond his own understanding near the end of the work. He describes guiding the collaboration toward a promising mathematical setting before the model advanced the argument. He also says the method resolves the generalized Riemann hypothesis, while a further extension remains prospective.
The release takes an unusual form. Alpöge says the formalization is finished, but that he intends to understand, simplify and explain the exceptionally long argument in a conventional mathematical write-up.
Lean is a proof assistant that can check whether a formal argument follows from specified definitions and assumptions. Its own documentation distinguishes two questions: whether a formal statement has a valid proof, and whether that statement means what its author intends.
For reviewers, that means examining the theorem’s definitions and dependencies as well as checking the proof itself. Lean’s validation guidance includes identifying any additional axioms or unfinished arguments in dependencies and, where appropriate, using independent checking tools. Publishing Lean files makes such scrutiny possible; it does not establish that every necessary check has already been completed.
A Much Bigger Claim Than Claude’s Earlier Result
The announcement follows Claude’s earlier advance on a related problem. In August, Anthropic said an unreleased research model had raised a lower bound for the fraction of zeta zeros on the critical line from 41.6% to 67.2%, drawing on existing mathematical research.
As we reported, that result did not prove the Riemann hypothesis. Establishing that a minimum proportion of zeros satisfies the prediction leaves open what happens to the rest. The new claim would close that gap, if validated.
AI-assisted mathematics has already produced other results requiring careful distinctions between a breakthrough and a complete solution. In the recent Yang-Mills review, physicists backed a proof addressing the existence of the theory, while the mass-gap requirement remained unresolved.
The Riemann hypothesis carries a $1 million Millennium Prize. A website announcement does not trigger an award: Clay’s rules require publication in a qualifying outlet, a wait of at least two years and general acceptance by the global mathematics community before it will consider a proposed solution.
Those requirements put independent scrutiny, rather than the announcement itself, at the center of what happens next.
Disclaimer: AIstify is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, technology, software, and digital innovation sectors. These relationships do not influence AIstify’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.