Levent Alpöge’s Unverified Riemann Claim Puts Claude’s Lean Code in Focus

Unite.AI
Read full story· 4 min
On November 20, 2026, mathematician Levent Alpöge announced a claimed proof of the Riemann hypothesis and a method for the Generalized Riemann Hypothesis, formalized as Lean code without a traditional write-up. Alpöge credits AI model Claude for key insights, though the proof awaits independent verification and further explanation.

What happened

On November 20, 2026, mathematician Levent Alpöge announced that he and the AI model Claude had produced a claimed proof of the Riemann hypothesis, along with a method for the Generalized Riemann Hypothesis. Alpöge released the argument as Lean code, a formal proof language, without a traditional written explanation. The proof has not yet been independently verified or reviewed by other mathematicians. Alpöge credited Claude for providing key mathematical insights during the process.

Key details

  • The Lean code was released in 14 compressed archive parts, with no conventional write-up.
  • Alpöge is a junior fellow at Harvard’s Society of Fellows and works at Anthropic.
  • The generated Lean document would exceed 250,000 pages if printed in full.
  • Claude previously helped increase the lower bound for zeta zeros satisfying the Riemann hypothesis from 41.6% to 67.2% in August 2026, but that was not a full proof.
  • No independent mathematical validation or peer review has yet been conducted on the new proof.

Why it matters

Mathematicians and researchers working on the Riemann hypothesis now have access to a formalized argument that claims to resolve one of mathematics’ most famous unsolved problems, but must wait for independent verification before it can be accepted. The use of AI model Claude as a collaborator introduces new questions about the role of AI in mathematical discovery and proof verification. Anyone seeking to build on or assess this work faces the immediate challenge of interpreting and checking a massive Lean codebase without a standard explanatory document.


More in AI Research