Riemann Hypothesis proof: Anthropic mathematician Levent Alpöge claims one with Claude, unverified so far
Reports differ on whether the claim covers the original Riemann Hypothesis or only the generalized version, and no independent mathematician has signed off.
Written with AI assistance from the sources listed below, then checked against them and reviewed by an editor. How our newsroom works
An Anthropic mathematician says he has found a Riemann Hypothesis proof with help from the company’s AI model Claude, but nobody outside the effort has confirmed it. Levent Alpöge announced the claim on Friday, 20 November, in a statement posted on his website, NBC Palm Springs reported. The 167-year-old problem is among the most famous unsolved questions in mathematics.
Riemann Hypothesis proof: what is claimed
“I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis,” Alpöge wrote, according to Tom’s Hardware. NBC said the proof also extends to the much broader Generalized Riemann Hypothesis. He released it with Lean files, a formal proof-checking system, rather than a traditional paper.
Interesting Engineering described a four-page explanation alongside a large archive of Lean verification files. It reported that an independent verification project has confirmed that the downloaded files match the published archive and that the main theorem statements compile against Mathlib, a mathematics library for Lean.
Generalized Riemann Hypothesis and the scope of the claim
Reports differ on what exactly is claimed. Interesting Engineering frames the result as a proof of the Generalized Riemann Hypothesis. NBC and Tom’s Hardware describe a proof of the Riemann Hypothesis itself, with the generalized version as an extension. Interesting Engineering said researchers still need to check that the code addresses the conventional Riemann Hypothesis, and to examine definitions and unsupported assumptions.
What has not been verified
NBC said the announcement “does not mean the problem has officially been solved” and that independent mathematicians still need to examine it. Citing The Verge, it said Anthropic has not independently endorsed the announcement. The company had earlier said it did not expect its techniques to lead to a proof. NBC also noted that Alpöge was previously associated with AI-assisted work on the Jacobian conjecture.
Key numbers
- $1 million: Clay Mathematics Institute prize for a verified solution, per NBC Palm Springs.
- 2 years: minimum scrutiny a solution must withstand under the prize rules, per the same report.
- 175,187: source files in the Lean archive, per Interesting Engineering.
- 239 MB: size of that archive, per the same report.
Clay Millennium Prize rules
The Riemann Hypothesis is one of the Clay Millennium Prize problems. NBC said a solution must gain broad acceptance and withstand at least two years of scrutiny before the $1 million is paid. Interesting Engineering added that Clay requires publication in an approved peer-reviewed outlet first. Even if checks go well, a payout is not near.
What happens next
Whether the Riemann Hypothesis proof holds depends on outside mathematicians checking both the written argument and the formal files. No timetable for that review has been reported.
A proof of the Riemann Hypothesis would settle a question that has stood since 1859 and would count among the biggest results in the field. The claim also tests how mathematicians verify AI-assisted work, since the argument comes with formal Lean files that outsiders can check but not yet a consensus.
Named in this story
Sources
- Anthropic Mathematician Says AI Helped Solve Math's Most Famous Unsolved Problem (NBC Palm Springs)
- Anthropic mathematician claims to have solved 167-year-old math problem with Claude (Tom's Hardware)
- Claude helped crack the 167-year-old Riemann hypothesis, Anthropic researcher claims (Interesting Engineering)