The Lean FRO and Mathlib Initiative have confirmed the correctness of Anthropic’s claimed solution to the Riemann hypothesis, a mathematical problem that had remained unresolved for 167 years.
In a joint statement on Saturday, the Lean and Mathlib communities said members had worked together over the previous 24 hours to independently examine the proof and establish its accuracy.
“We hereby certify that both the statement and the proof of the Riemann Hypothesis are correct,” the statement said.
On Friday, Anthropic mathematician Levent Alpöge announced on his website that he had used Claude artificial intelligence (AI) to solve the longstanding mathematical problem.
The hypothesis states that the real part of every non-trivial zero of the Riemann zeta function is equal to one-half.
The Lean FRO and Mathlib Initiative said it assembled a team of experts who spent the day examining Alpöge’s mathematical argument before confirming its validity.
The groups expressed confidence in the findings and congratulated the Anthropic mathematician on the achievement.
Alpöge had explained that he formalised the solution using Lean, a computer programming language, with Claude AI contributing approximately 250,000 pages of work towards the proof.
According to the initiative, the verification process involved four criteria used to assess Alpöge’s 17 million lines of code.
The Riemann hypothesis was listed in 2000 as one of the seven Millennium Prize Problems, a collection of some of the most challenging unsolved mathematical questions, each carrying a $1 million prize.
The verified proof represents the second Millennium Prize Problem reportedly solved with the assistance of artificial intelligence, following OpenAI’s solution to the Navier-Stokes existence and smoothness problem in September.
