Anthropic Mathematician Claims Claude AI Helped Solve the 167-Year-Old Riemann Hypothesis

Anthropic Mathematician Claims Claude AI Helped Solve the 167-Year-Old Riemann Hypothesis

Anthropic mathematician Levent Alpöge announced on November 20, 2026, that he and the company's Claude AI model may have developed a proof of the Riemann hypothesis, one of mathematics' most difficult unsolved problems. Alpöge released the proposed solution as computer-verifiable code rather than a traditional research paper, claiming the work also establishes the Generalized Riemann Hypothesis. The findings have not yet undergone independent verification.

The Riemann hypothesis, introduced in 1859, concerns the mathematical patterns governing the distribution of prime numbers. It has remained unresolved for 167 years and is among the seven Millennium Prize Problems recognized by the Clay Mathematics Institute, each carrying a $1 million award.

Alpöge, who also serves as a junior fellow at Harvard's Society of Fellows, published a brief statement on his personal website outlining the claimed discovery and Claude's involvement in developing the argument.

"I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis," Alpöge wrote.

The proposed solution was developed using Lean, a mathematical proof assistant that allows researchers to encode arguments in a form that software can check for logical consistency. Unlike a conventional paper, which requires mathematicians to evaluate written reasoning, a Lean formalization is designed to verify individual steps against precisely defined mathematical statements.

Alpöge distributed the work across 14 compressed archives containing approximately 175,187 source files and totaling 239 MB. He said the formalization contains no unproven assumptions, although outside researchers have not yet confirmed that the complete code verifies successfully or establishes the intended mathematical result.

The unusually large submission also explains why Alpöge did not prepare a conventional paper. He estimated that translating the argument into a written mathematical presentation would require more than 250,000 pages, making a traditional publication impractical in its current form.

According to Alpöge, the approach adapts Stepanov's method, a mathematical technique previously used in the function-field setting. A central challenge involved reproducing a particular mathematical effect over the integers, which the proposed argument addresses through repeated theta lifts within a framework involving automorphic forms.

Alpöge credited Claude with helping determine the sequence of those mathematical operations. He described his own role as directing the broader strategy while the AI contributed specialized knowledge and accelerated the later stages of the work.

A development timeline included in his statement places an important determinant calculation on October 6, followed by progress involving repeated theta lifts on November 15. The extension to the Generalized Riemann Hypothesis reportedly followed on November 16.

That broader result would extend the original conjecture to Dirichlet L-functions, another important class of mathematical functions. Alpöge also suggested that the approach might eventually address the Extended Riemann Hypothesis, but he has not claimed to have established that result.

The announcement follows earlier Anthropic research in which Claude contributed to progress on the same mathematical problem without producing a complete proof.

In August, Anthropic reported that an experimental version of Claude had improved the established lower bound for the proportion of relevant zeta-function zeros satisfying the hypothesis from 41.6% to 67.2%.

That research involved two Claude Code sessions, approximately 31 million generated tokens and around 60 coordinating AI subagents. Anthropic explicitly stated that the work did not resolve the Riemann hypothesis.

The earlier findings were examined by Alpöge and mathematician Ralph Furman, with additional review from number theorists Brian Conrey and Dan Goldston. No equivalent independent assessment has been announced for Alpöge's latest submission.

The new claim has nevertheless attracted attention from established mathematicians. James Maynard, a University of Oxford mathematics professor and 2022 Fields Medal recipient, told IBM Think that the result would represent a major achievement if verified.

"If this fully checks out, it is clearly a historic milestone in mathematics, establishing one of the great mysteries of the subject," Maynard said.

He also expressed cautious optimism about the proposed proof's validity, while acknowledging that examining the formal verification could require considerable time.

"I suspect that the proof will turn out to be correct, although even checking the Lean verification will take some time," Maynard said.

The verification process presents challenges beyond simply running the code. Mathematicians must establish that the formal definitions accurately represent the conventional Riemann hypothesis and that the entire argument can be checked without relying on unsupported assumptions.

Researchers would also need to understand the mathematical reasoning behind the result. A successful computer verification could establish the correctness of a formal statement without necessarily providing an accessible explanation of the techniques that made the proof possible.

Alpöge himself acknowledged uncertainties surrounding the work, including the possibility that its construction contains unnecessarily complicated mathematical choices. He has not presented the announcement as a result already accepted by the mathematical community.

Anthropic has not issued an official statement confirming the discovery. Although Alpöge thanked several mathematicians, including Manjul Bhargava and Peter Sarnak, his announcement did not identify them as independent reviewers of the proposed proof.

The Clay Mathematics Institute also maintains requirements that would prevent an immediate prize award. A qualifying solution must be published in an appropriate peer-reviewed venue and withstand at least two years of examination before becoming eligible for the $1 million prize.

For now, Alpöge's announcement remains an extraordinary but unverified mathematical claim. If the Lean formalization withstands independent scrutiny, it could establish a solution to a problem that has resisted mathematicians for more than a century and a half, while demonstrating a significant role for AI in developing advanced mathematical proofs.

This analysis is based on reporting from IBM.

Image courtesy of Anthropic.

This article was generated with AI assistance and reviewed for accuracy and quality.

Updated Nov 20, 2026

About this article: This article was generated with AI assistance and reviewed by our editorial team to ensure it follows our editorial standards for accuracy and independence. We maintain strict fact-checking protocols and cite all sources.

Word count: 924Reading time: 0 minutes

AI News Daily

Breaking Intelligence • Since 2023

Join hundreds of thousands of AI professionals who start their day with our curated newsletter. Get breaking news, expert analysis, and exclusive insights.

Stay Ahead of AI

Get the latest AI breakthroughs, tools, and insights delivered to your inbox every week.

✓ Free forever✓ Unsubscribe anytime✓ No spam guarantee

Go Premium

Unlock unlimited AI tools and an ad-free reading experience designed for AI professionals.

• Ad-free experience• Premium AI tools
Start Free Trial

14-day free trial • Cancel anytime
Plus $9/mo • Pro $90/yr (2 months free)

Follow Our Community

ChatAI

Breaking Intelligence

Your daily briefing on what matters in AI. Trusted by developers, researchers, executives, and AI enthusiasts worldwide.

© 2026 ChatAI. All rights reserved.