News

Riemann hypothesis falls to Anthropic mathematician and Claude in six weeks

What's the deal? On November 20, 2026, mathematician Levent Alpöge released a proof of the Riemann hypothesis as a Lean library of 17 million lines of code. A day later, the Lean and Mathlib communities said they have "full confidence" the proof is correct.

Why it matters: The Riemann hypothesis is one of mathematics' oldest and most famous open problems. Alpöge said he produced the proof by working with Anthropic's ClaudeDealroom has a profile for this one. Try Dealroom → over roughly six weeks.

How they checked it: The communities certified both the statement and the proof against four published criteria. Dozens of people reproduced the same proof export, two independent tools confirmed the theorem matches Mathlib's statement of the hypothesis, and several proof checkers — including Lean's official kernel — validated the result.

Two of those checkers, con-leche and con-ron, come with a formal proof that they cannot accept a false proof, addressing concerns that Lean's own logic might be exploited. The checks ran on an inert export, out of reach of any malicious code in the proof sources.

The catch: Certification confirms the proof is valid, not that anyone yet understands it. Mathematicians still have to digest the new ideas. "We look forward to all the activity in the mathematical community that will help explain the new ideas in this proof," the communities wrote.

The signal: This is arguably the AI-for-science story of the year, landing during Anthropic's IPO week. A human-plus-model collaboration cracking a landmark problem in six weeks — and a community verifying it in a day — points to where machine-assisted mathematics is headed.

Read more: leanprover-community.github.io

Image credit: Generated with Gemini

More top stories