researchAnthropic

Anthropic mathematician Levent Alpöge claims Claude-assisted proof of the Riemann hypothesis, posts it in Lean

TL;DR

Levent Alpöge, a mathematician employed by Anthropic, says he has proven the Riemann hypothesis, first proposed in 1859, working with Claude. He posted the proof coded in Lean, and experts have yet to confirm that it is correct.

3 min read
0

Levent Alpöge, a mathematician and employee at Anthropic, claimed on Friday to have proven the Riemann hypothesis using Claude. He posted the proof, coded in the Lean programming language, on his website. Scientific American, which reported the claim, describes it as a breaking story and says it will be updated.

"I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis," Alpöge wrote in a statement on his website.

What is claimed

  • The problem: The Riemann hypothesis was first proposed in 1859 by Bernhard Riemann. It holds that the locations of the prime numbers along the number line, which appear random at first, are determined by an intricate, meaningful order. Scientific American calls it the most famous open problem in mathematics and the "North Star" of number theory.
  • The scope: Alpöge says his proof confirms the Generalized Riemann Hypothesis, a much broader expansion of Riemann's original guess.
  • The format: The proof is posted in Lean, which verifies the mathematical correctness of each argument.

Verification status

The claim has not yet been confirmed. Scientific American reports that issues can arise in Lean, such as the hypothesis being coded incorrectly, but that it should not take long for experts to check whether the statement is properly formalized and whether the proof is correct.

Background

Scientific American reports that the first sign of progress came in October, when one of the proofs in OpenAI's large release of math results made significant progress on making sense of the primes. Alpöge has been in the headlines since July, when he tweeted a Claude-generated solution to the Jacobian conjecture. Scientific American says that post kicked off an unprecedented summer in mathematics.

The outlet frames the claim as the capstone of a year in which AI's mathematical capabilities rose from "unreliable college student to a superhuman force." That characterization is Scientific American's.

Alpöge's statement acknowledges the grief some mathematicians feel. "It is natural to feel that this is an end, in my case a loss of a principal guiding light," he wrote. "But what I saw during the course of progress so amazed me that I believe it is one dawn across science of something far greater." He added: "I hope the argument provides a glimpse towards the miracles that lie ahead."

The report does not state which Claude model was used.

What this means

The immediate question is whether the proof holds. Because it is written in Lean, checking it is a matter of confirming that the formal statement matches the Riemann hypothesis and its generalized form, and that the proof compiles. Per Scientific American, that review should be quick for experts.

If the proof is confirmed, Scientific American says the Generalized Riemann Hypothesis would instantly unlock a large set of results that currently depend on it. Until that review is complete, this remains Alpöge's claim, and he himself writes "I believe." Alpöge is an Anthropic employee and the proof was produced with Claude, so independent verification by outside mathematicians will determine how the result is received.

Related Articles

research

AWS releases I/Q, I/R and Isabelle Assistant to connect AI agents to the Isabelle proof checker via MCP

AWS has made three tools in its AutoCorrode verification suite publicly available: I/Q, I/R and Isabelle Assistant. They connect AI agents to the Isabelle proof assistant through the Model Context Protocol, so generated proofs are checked by computer. Amazon Science authors demonstrate the setup using Claude Opus 4.6.

product update

Anthropic Adds Experimental Code Mode to Claude Code v2.1.323, Chaining Tool Calls in JavaScript

Anthropic has added an experimental Code Mode to Claude Code v2.1.323. Claude can write JavaScript that chains Bash, Read and Edit calls in one step, and keep talking while long commands run.

changelog

Anthropic launches Linked billing beta, letting Claude API developers bill usage to a customer's own organization

Anthropic launched Linked billing in beta on November 17, 2026. Developers can bill Claude API usage in their applications to a customer's own Claude organization, choosing per request whether the customer or the developer pays.

benchmark

Vals tests: Claude Opus 5.5 picked Minab school as military target in 22 of 32 runs, Bloomberg reports

AI models from eight companies often marked civilians, ambulances, surrendering soldiers and a school in Iran as military targets in simulated tests, Bloomberg reported, citing a report by evaluator Vals. Anthropic's models did worst, while OpenAI's were among the best but still made mistakes.

Comments

Loading...