Anthropic mathematician claims AI model Claude helped prove Riemann Hypothesis
An Anthropic mathematician has published a note claiming to have used the Claude model to prove the Riemann hypothesis.
An Anthropic mathematician has published a note claiming to have used the Claude model to prove a major mathematical conundrum, raising questions across the scientific community.
Levent Alpöge, a mathematician employed by Anthropic, announced the claim in a note published on his personal website on Nov. 20, 2026. The announcement arrives on the heels of the company's recent stock market debut, prompting some observers to note the coincidence in timing, though Alpöge dismissed the cynical view in his public statements. The claim centers on the Riemann hypothesis, originally proposed in 1859 by German mathematician Bernhard Riemann. The hypothesis concerns the zeros of the Riemann zeta function and their connection to the distribution of prime numbers, which are divisible only by themselves and one.
Media additions
According to the four-page announcement titled "Remark," Alpöge stated that he believes he has, working with Claude, discovered a proof of the Riemann Hypothesis. He added that the methodology extends to the Generalized Riemann Hypothesis and potentially the Extended Riemann Hypothesis given more time. The announcement describes a collaborative process where Alpöge provided high-level direction while the model contributed significantly near the conclusion. "Claude far outpaced me near the finish line," Alpöge wrote, adding that the model "flew, and danced, and sparkled near that finish line."
The problem is considered a central challenge in number theory and forms one of the Millennium Prize Problems designated by the Clay Mathematics Institute, which offers a $1 million reward for an accepted solution. However, independent experts caution that an online announcement and accompanying computer files do not constitute official verification.
Rather than distributing a conventional academic manuscript, Alpöge released his argument formalized in Lean, a programming language and proof assistant that allows computers to check logical steps. Independent researchers, including MIT mathematician Andrew Sutherland, have separately compiled the Lean files without encountering technical errors. Yet, computer compilation alone does not guarantee that the code faithfully represents the mathematical statement or avoids unintended shortcuts.
Prominent mathematicians have expressed a mixture of awe and caution regarding the announcement. James Maynard, the Regius Professor of Mathematics at the University of Oxford and a Fields Medal winner, told IBM Think that if the proof holds up, it represents a historic milestone. Nevertheless, Maynard noted that verifying the computer checks and understanding the underlying ideas will require substantial effort. Youness Lamzouri, a mathematics professor at the University of Lorraine, voiced astonishment at the claim while emphasizing that mathematical verification requires human comprehension of why the arguments work.
The announcement follows a broader trend of rapid advancement in automated reasoning systems. Over recent years, artificial intelligence models have progressed from struggling with basic arithmetic to winning medals at international competitions and tackling complex open problems across various scientific disciplines. Earlier milestones included work on the Navier-Stokes equation and alterations to lower bounds for the Riemann zeta function on the critical line.
| Milestone / Event | Approximate Date / Year | Context |
|---|---|---|
| Initial Hypothesis Proposal | 1859 | Bernhard Riemann outlines zeta function zeros and prime number distributions. |
| Zeta Function Bound Improvement | August 2026 | Anthropic reports a research Claude model improves known lower bounds on the critical line. |
| Alpöge's Proof Announcement | Nov. 20, 2026 | Mathematician publishes Lean formalization claiming proof of Riemann and Generalized Riemann hypotheses. |
Reactions outside academia reflect shifting perspectives on intellectual labor in the digital age. Danning Sui, a research director at Pantera Capital, suggested on social media that frontier scientific discovery risks resembling calculator usage, potentially altering how subjects like mathematics are taught and pursued. Meanwhile, Alpöge expressed a personal sense of loss alongside his wonder, describing the Riemann hypothesis as a childhood guiding light.
What to Watch Next
- Independent researchers will continue scrutinizing the Lean formalization line by line to verify compilation and logical soundness.
- Experts will work to translate the computer-checked logic into human-readable explanations to ensure the theorem matches the classical problem.
- The Clay Mathematics Institute maintains strict rules for its Millennium Prize, requiring formal publication in a qualifying journal and a minimum two-year review period before any official acceptance.
For more updates on corporate developments, see our ongoing Business coverage.