❯ Anthropic mathematician says he and Claude found a proof of the Riemann hypothesis and releases Lean code for review
From a simpler case to the original problemOn November 20, Levent Alpöge released a note saying he and Claude had found a proof of the Riemann hypothesis and that the method extended to the generalized hypothesis. His timeline runs from October 6 to mid-November. He began with a simplified version, joined two previously separate lines of reasoning and steered Claude toward the original problem. He says Claude took on crucial technical work near the end in areas beyond his own expertise.
What the August result establishedThe hypothesis asks whether all nontrivial zeros of the zeta function lie on one critical line. Their location constrains the error in estimates of how primes are distributed. In August, Anthropic published a separate result in which a research version of Claude raised the established lower bound on the proportion of zeros on that line from 41.6% to 67.2%. The team did not expect its method to yield a direct proof for every zero.
How reviewers can check the releaseAlpöge released Lean code in 14 archives and plans to simplify the argument and write a paper. Outside researchers need to build the project independently, inspect whether the final theorem faithfully states the conjecture and check whether the argument relies only on the stated standard axioms. The claim and materials are public. Independent verification and a formal paper remain pending.
From a code release to Millennium Prize considerationRiemann proposed the hypothesis in 1859. The Clay Mathematics Institute later named it one of seven Millennium Prize Problems. Clay considers a solution only after publication in a qualifying outlet, at least two years have passed and the international mathematics community has generally accepted it. Independent researchers can check the code much sooner, on a separate timetable.
▮ SIGNALThe Lean release gives outside mathematicians a way to check the theorem statement and its axioms before assessing the argument developed with Claude.