Anthropic researcher claims Claude helped him prove the Riemann hypothesis

What happened
Levent Alpöge, a mathematician at Anthropic, announced on his website that he has used the Claude AI model to prove the Riemann hypothesis, a longstanding unsolved problem in mathematics. He released a 17-million-line proof written in the Lean programming language, which checks each logical step. Several mathematicians, including Andrew Sutherland of MIT, have run independent checks on the code and found no errors. However, experts have not yet formally accepted the proof, and some caution that human understanding and verification are still needed.
Key details
- Levent Alpöge is a researcher at Anthropic.
- The proof was written in Lean and is about 17 million lines long.
- The Riemann hypothesis is one of the Clay Mathematics Institute’s Millennium Prize Problems, with a $1 million prize.
- Andrew Sutherland and at least three other mathematicians have checked the code independently.
- The proof also claims to address the generalised Riemann hypothesis.
Why it matters
Mathematicians and the broader mathematics community are directly affected, as a solution to the Riemann hypothesis would resolve a major open question and potentially award a $1 million prize. Researchers and institutions now face the task of formally verifying the proof and ensuring it addresses the correct problem. The use of AI models like Claude in this process introduces new methods for tackling complex mathematical challenges, but also requires additional scrutiny to confirm results.
- Who is Levent Alpöge? Meet Anthropic’s member who cracked Riemann hypothesis - the world’s most famous maths problem· Mint
- Two separate maths groups confirm Anthropic’s Levent Alpöge solved the Riemann Hypothesis | Here’s what they said· 2 sources
- Claude solves mathematical problem that remained unsolved for 160 years: What is Riemann hypothesis?· 2 sources
- In a year of AI solving maths problems, an Anthropic worker says he may have cracked the biggest· 20 sources
