i
News
News · 2026-11-21

Two organizations confirm Claude-assisted Riemann hypothesis proof

Dataist Dataist

Two organizations have confirmed a mathematician’s claim: Claude helped prove a mathematical conjecture.

Cover: Two organizations confirm Claude-assisted Riemann hypothesis proof

Anthropic mathematician Levent Alpöge said Claude helped him prove the Riemann hypothesis, one of mathematics’ most important unsolved problems. The work took about six weeks. On November 18, two organizations that work with the Lean programming language independently checked the proof and confirmed it was correct. Bernhard Riemann formulated the hypothesis in 1859; the Clay Mathematics Institute has offered a $1 million prize for solving it.

Claude helped a mathematician prove the Riemann hypothesis

The proof was checked in Lean

On Friday, Alpöge posted a message on his personal website and included code in the Lean language that, he said, could be used to check the proof. The next day, Lean Focused Research Organization, which develops the use of Lean, and Mathlib Initiative, which uses the language for mathematical equations, said they had independently confirmed the result.

The organizations said they were fully confident in the correctness of Alpöge’s proof and congratulated him on the achievement. Their check concerned the solution he presented, written in Lean.

Alpöge is a Turkish-American mathematician. He wrote that he and Claude had worked on a wide range of problems in recent months, but that in the final stretch the system had significantly outpaced him. In his post, he described Claude as if it could fly and dance.

The announcement came just one day after Anthropic went public on Nasdaq. Alpöge called the timing unfortunate and wrote that he would have finished earlier if he could.

What the Riemann hypothesis is

The Riemann hypothesis is a problem that is nearly 170 years old. German mathematician Bernhard Riemann proposed it in 1859. It concerns prime numbers: integers greater than 1 that are divisible only by themselves and 1.

Riemann studied how often prime numbers occur and how they are distributed along the number line. He also developed a formula for estimating the number of primes below a given number.

The hypothesis is one of the seven Millennium Prize Problems set by the Clay Mathematics Institute. The institute offers $1 million for solving each one.

Disputes over the use of AI in mathematics

Mathematicians have tried to solve the problem for years. The New York Times reported that on Friday, independent Italian mathematician Alessandro Mazzullo published an unfinished proof of the hypothesis. He had worked on it for nearly 12 years and noted that he had not used AI because he did not fully trust such systems.

Debate over AI’s role intensified after OpenAI said in September that a closed version of its tools had helped solve part of another Millennium Prize Problem. The method drew bewilderment from mathematicians and observers: a private company valued at nearly $1 trillion spent $15 million and used 10,000 AI systems. At the time, Alpöge was working on the same separate problem for Anthropic.

In September, University of Warwick mathematician James Robinson told The Guardian that he was annoyed when large technology companies burned resources to show that their new models were superior. He compared the behavior to immature playground boasting and added that billions of dollars and the risk of serious environmental harm lay behind it, against the backdrop of climate change, which he called probably the biggest problem of our time.

Dataist's take: Checking the proof in Lean makes it possible to independently verify a result claimed with AI’s involvement. But disputes over the cost and environmental impact of such computing remain part of a broader question: do mathematical achievements justify the resources technology companies spend?

Daily AI news

Every day we pick what actually matters in AI and explain it plainly — no hype, no filler. Subscribe if you want to follow where the industry is going.

Only what matters — every day

Follow on X