What happened
Levent Alpöge, a mathematician employed by Anthropic, posted a note on his personal website asserting that the company’s AI system Claude contributed to a proof of the Riemann hypothesis. He provided a link to Lean code that he says can verify the proof and noted that the work took roughly six weeks. The claim was made public a day after Anthropic’s Nasdaq IPO and has prompted other mathematicians to attempt independent verification.
On 21 November 2026, Levent Alpöge posted a brief note on his website claiming that Claude, Anthropic’s flagship conversational AI, helped him develop a proof of the Riemann hypothesis. Alpöge described the collaboration as lasting about six weeks and said Claude "far outpaced" him near the finish line.
Alpöge attached a repository of Lean code—a formal proof language used by mathematicians—to his note, asserting that the code can be used to verify the claimed result. He did not provide a peer‑reviewed paper or a formal announcement from a mathematical society.
The claim emerged one day after Anthropic’s initial public offering on the Nasdaq, a timing Alpöge himself described as "horrible" and suggested he would have preferred to release the work earlier.
Other mathematicians, including University of Warwick’s James Robinson, have expressed skepticism about AI‑generated proofs, citing concerns over reproducibility, environmental impact, and the influence of large private funding on academic research.
The Guardian report notes that a separate, unrelated claim of an AI‑assisted partial solution to another Millennium problem was made by OpenAI in September, further underscoring the competitive landscape of AI‑driven mathematics.
Source details: theguardian.com
Why it matters
If confirmed, a proof of the Riemann hypothesis would resolve one of the seven Millennium Prize problems, a set of the most challenging open questions in mathematics, and would be the first such breakthrough attributed to an AI‑assisted effort. The claim highlights the growing role of large language models in high‑level mathematical research, raising questions about the credibility, reproducibility, and ethical implications of private‑sector AI contributions to fundamental science. It also fuels ongoing debate about the environmental and financial costs of deploying massive AI resources for academic pursuits.
The Riemann hypothesis, formulated in 1859, underpins much of analytic number theory and has implications for the distribution of prime numbers. A valid proof would resolve a problem that has resisted solution for nearly two centuries and would earn a $1 million Clay Mathematics Institute prize.
An AI‑assisted proof would be a landmark demonstration of large language models extending beyond language tasks into rigorous, formal mathematics, potentially reshaping how complex proofs are approached and verified.
The claim raises transparency issues: Anthropic has not disclosed whether Claude was accessed via a public API, a private research license, or an internal tool, nor has the company commented on the claim. This opacity makes it difficult for independent researchers to replicate or evaluate the methodology.
Environmental and financial concerns are amplified by prior reports that OpenAI spent $15 million on 10,000 AI instances to tackle a different Millennium problem. If Claude’s involvement required comparable , the claim could intensify scrutiny of the carbon footprint of AI‑driven scientific research.
The timing of the announcement—immediately after Anthropic’s IPO—could influence investor perception of the company’s technological edge, potentially affecting market dynamics in the AI sector.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
Which component of an AI application is the machine-learning model itself?
What to watch next
Mathematicians will be scrutinizing the Lean code and any accompanying formal verification to determine whether the proof holds up under peer review. The broader AI community will watch how Anthropic and other firms respond to the claim, including any statements about access to Claude for academic use. Finally, the episode may influence policy discussions on transparency and funding for AI‑driven research, especially given the timing around Anthropic’s IPO.
Formal verification: Mathematicians will attempt to run the provided Lean code, check its correctness, and assess whether it indeed constitutes a complete proof of the hypothesis.
Peer review: Publication of the proof in a reputable mathematical journal or endorsement by a recognized body such as the Clay Mathematics Institute would be needed for the claim to gain acceptance.
Anthropic’s response: Any official statement from Anthropic regarding Claude’s role, access policies for researchers, or plans to make the tool available for academic work will be closely monitored.
Regulatory and policy reactions: Lawmakers and funding agencies may revisit guidelines on AI transparency and the disclosure of AI contributions to scientific breakthroughs.
Market impact: Investors may watch Anthropic’s stock for volatility linked to the claim’s verification status, especially given the proximity to the company’s recent IPO.