AIGenerated
EN · IT · ES

Anthropic worker says Claude helped prove the Riemann Hypothesis, a 167-year-old math mystery

If confirmed, it would mark the first time an AI has solved one of mathematics' most famous unsolved problems.

Pisphere Solution representative of P Versus NP and Riemann Hypothesis.
Photo: Enjoyclear, CC BY-SA 3.0, via Wikimedia Commons (licence)
How it works

A computer program that helps people solve hard math problems.

The claim

Levent Alpöge, a researcher who works for Anthropic, announced that he spent six weeks building a detailed proof of the Riemann Hypothesis together with Claude 1. He described the problem as "the sacred object in mathematics" for him since childhood .

The complete proof was written in Lean, a programming language designed to let people write and automatically check mathematical arguments 1. Alpöge said the finished document runs to more than 250,000 pages long .

What the Riemann Hypothesis is about

In 1859, Bernhard Riemann proposed that prime numbers — whole numbers like 2, 3, 5 and 7 that cannot be divided evenly by anything other than themselves and one — follow a hidden pattern along the number line 1. The hypothesis tries to describe exactly where those primes appear .

It is widely seen as the most important open question in all of mathematics. The Clay Mathematics Institute counts it as one of its Millennium Prize Problems and offers $1 million for the first verified solution 1.

At a glance

How a 167-year-old math mystery reached today

  1. 1859Bernhard Riemann proposes his guess about where prime numbers appear
  2. OctoberOpenAI shares hundreds of math solutions and advances toward the hypothesis
  3. Six weeks agoAnthropic researcher Levent Alpöge starts building a proof with Claude

Dates from the story; the proof is still awaiting full independent verification.

How certain are experts?

No independent confirmation has come yet 1. James Maynard, a mathematician at the University of Oxford, told the New York Times he was "98 or 99 percent certain" the proof is correct . That kind of high confidence from a peer gives the claim extra weight, but it still falls short of full verification .

Alpöge's work is not the first time an AI model has been linked to progress on this problem. In October, OpenAI posted hundreds of solutions to past math challenges and shared advances toward the Riemann Hypothesis itself 1. Those results also need time to be checked by the math community .

If researchers eventually verify Alpöge's argument, it would mean an AI system helped solve a puzzle that even the brightest human minds could not crack over 167 years 1.

Words to know

Prime number: a whole number bigger than 1 that can only be divided evenly by 1 and itself 1. · Riemann Hypothesis: a guess about how prime numbers are arranged, proposed in 1859 and still unproven . · Lean: a computer program that checks whether a mathematical proof is logically correct .

Key facts

  • Anthropic researcher Levent Alpöge claims Claude helped build a proof of the Riemann Hypothesis
  • The proof is written in Lean and exceeds 250,000 pages
  • Oxford mathematician James Maynard says he is '98 or 99 percent certain' the proof is accurate
  • The Riemann Hypothesis has been unsolved since 1859 and carries a $1 million prize
  • OpenAI also shared advances on the problem in October; neither result is yet independently verified

Sources

  1. Anthropic employee claims to have proven the most famous math problem using Claude - Engadget — engadget.com

If this story concerns you or your company, or uses something that belongs to you, ask for a correction or for it to be taken down. How requests work

Keep reading