Book a call

News · November 20, 2026

A mathematician says he proved the Riemann Hypothesis with Claude

// ai.news :: 2026

Anthropic mathematician Levent Alpöge has posted a note claiming that he and Claude found a proof of the Riemann Hypothesis, released as a Lean formalization rather than a paper. Nobody has checked it yet. If it holds, it would be the clearest sign so far that AI can do research-level work, not just assist with it.

What happened

  • A four-page note posted on Alpöge's personal website on November 20, 2026 says that, working with Claude, the author found a proof of the Riemann Hypothesis, and that the same method also proves the Generalized Riemann Hypothesis, IBM Think and OfficeChai reported.
  • According to the note, the argument has been formalized in the Lean proof assistant without extra assumptions, and only the Lean files are being released for now because the informal argument is too long to present; a readable write-up is promised later.
  • The note credits Claude with the key step of the argument and with the parts dealing with higher-rank automorphic forms, and says that toward the end Claude moved far faster than the author; it does not name a specific Claude model.
  • The proof has not been independently verified, IBM Think reported, and checking the Lean files will take time; OfficeChai noted that no one outside the author's circle has yet confirmed that the archive builds or that its final theorem states the Riemann Hypothesis.
  • IBM Think reported that Fields Medalist James Maynard of Oxford called the result potentially historic if it checks out and said he suspects it will prove correct, while Youness Lamzouri of the University of Lorraine said mathematicians may need months or years to understand it.

Why it matters

We do not usually cover pure mathematics, and this is an exception. CMM's founders and leadership come from applied mathematics, and the Riemann Hypothesis, posed in 1859, is the problem every mathematician knows. A claim like this, made with a commercial AI model, deserves attention even before it is checked. The honest status today is simple: a claim, not a result. Formal verification in Lean is the right way to settle it, and the community will now do that work.

The more durable lesson for business is the shape of the workflow, not the headline. A human expert set the direction, the model did a large share of the technical work, and the output was delivered in a form a machine can check. That is the same pattern that works in software: AI writes, tests and types verify. Teams that invest in strong automated checks, such as test suites, type systems, static analysis and formal specs where they matter, get far more safe value out of AI than teams that rely on reading the output.

There is also a warning in the details. The note says its own informal explanation may not make sense and that the full argument is not yet readable by humans. Correct but opaque work is hard to maintain, extend or trust. In production code, insist that AI-generated changes come with explanations a reviewer can follow, not just a green build.

Sources

  • Levent Alpöge (personal website): Remark
  • IBM Think: Anthropic mathematician claims AI helped solve a 167-year-old math mystery
  • Scientific American: Anthropic mathematician claims to prove Riemann hypothesis, most famous problem in all of math
  • OfficeChai: Claimed Riemann Hypothesis Proof Made With Claude Appears On Anthropic Mathematician Levent Alpöge's Website

This post is CMM’s own summary and commentary on publicly reported news. Facts are attributed to the sources linked above; please read them for full details. CMM is not affiliated with these publishers or with the companies mentioned, and trademarks belong to their owners. Prepared with AI assistance. Corrections or requests: [email protected]. See our editorial policy.

Want help with this?

Tell us what you need, and we will suggest the right team model on a short call.

Talk to us