Anthropic may have reached a historic milestone in mathematical research. On November 20, 2026, Levent Alpöge, a mathematician working for the company, announced he had discovered with Claude
Anthropic may have reached a historic milestone in mathematical research. On November 20, 2026, Levent Alpöge, a mathematician working for the company, announced he had discovered with Claude a proof of the Riemann hypothesis, formulated in 1859. This problem, unsolved for 167 years, is associated with a one million dollar reward. The announcement comes the day after Anthropic’s record-breaking IPO. However, the proof still needs to be reviewed by independent experts.
In Brief
- An Anthropic researcher claims to have proved the Riemann hypothesis with Claude.
- The proof notably relies on 17 million lines of mathematical code.
- The announcement comes one day after Anthropic’s historic IPO, which raised 81 billion dollars.
Anthropic announces a potential historic mathematical breakthrough
Levent Alpöge believes he has solved one of the most famous problems in mathematics. On Friday, November 20, this Anthropic researcher published a four-page paper on his personal website. He claims to have worked for several weeks with Claude to construct a proof of the Riemann hypothesis.
The artificial intelligence had previously solved a 25-year-old cryptographic mystery related to the NSA. This time, the challenge is significantly greater. Formulated in 1859 by the German mathematician Bernhard Riemann, the hypothesis concerns the distribution of prime numbers.
These numbers, such as 2, 3, 5, or 7, are only divisible by themselves and 1. Their distribution follows complex mathematical properties that researchers have been trying to understand for centuries.
Riemann established a link between this distribution and a particular mathematical function, called the zeta function. His hypothesis states that all non-trivial zeros of this function have a real part equal to 1/2.
The statement seems brief. Yet its proof has resisted the best mathematicians for 167 years. Alpöge claims to have gone further. According to him, the method developed with Claude would also prove the generalized Riemann hypothesis, a version extending the problem to other mathematical functions. The researcher acknowledges that Claude surpassed him in some final stages of the reasoning. He has published the technical elements of his proof to allow for their examination.
Secure your cryptos with SafePalThis link uses an affiliate program.
A 17-million-line Claude proof to verify
The announcement is based on particularly large-scale computational work. Alpöge distributed about 17 million lines of code in Lean, a language used to formalize and verify mathematical proofs.
The researcher also explains having generated over 250,000 pages of documents during the process. Lean allows translating reasoning into instructions submitted to rigorous logical rules. Its use provides an additional way to check certain steps of a proof.
But the work still needs to be independently reviewed. Recent advances in these tools had already caused concern among top mathematicians, notably about the future role of human researchers in scientific discovery.
The Clay Mathematics Institute still classifies the Riemann hypothesis among unresolved problems. The institute selected seven major mathematical challenges in 2000, with a one million dollar reward for each.
The awarding rules are strict. A proof must be published in an admissible scientific publication, gain general recognition from the mathematical community, and undergo a period of at least two years after publication.
An announcement published on a personal website is therefore insufficient. Martin Bridson, president of the Clay Mathematics Institute, believes this statement could have historical significance, while emphasizing that understanding and evaluation will take time. For Alpöge, much work remains. The researcher admits he has not yet personally mastered all the mathematical fields involved in this immense proof.
Anthropic reveals this discovery just after its IPO
The timing is especially notable. On November 19, Anthropic completed the largest IPO in history, raising about 81 billion dollars for a valuation close to 1.5 trillion.
The next day, one of its researchers announces a possible proof of the Riemann hypothesis with Claude. Alpöge himself acknowledges the delicate nature of this coincidence. In his paper, he explains having obtained his decisive results in the night of Sunday, November 15. He then contacted Dario Amodei, CEO of Anthropic, to discuss conditions for a public announcement.
The researcher rejects any interpretation that he chose this date to serve the company’s financial interests. No public evidence demonstrates manipulation of the timeline for stock market purposes. However, this announcement occurs during a period of strong investor attention towards Anthropic.
Even before the IPO, crypto traders were betting on a valuation above the official price of 83 dollars per share, notably through derivative contracts traded on Hyperliquid. A recognized mathematical discovery could strengthen Claude’s scientific reputation. It would also provide a major example of the use of artificial intelligence in fundamental research.
However, it would be premature to infer a quantifiable commercial advantage or a future rise in Anthropic’s stock. For now, the company has a considerable scientific announcement, accompanied by a proof that specialists must review. If it withstands this review, Claude will have contributed to solving a problem that has occupied several generations of mathematicians.