A mathematician employed by Anthropic asserts that the San Francisco-based AI firm’s software has played a pivotal role in solving one of the most significant unsolved challenges in mathematics.
Levent Alpöge, a Turkish American mathematician, revealed on his personal website that Claude helped him develop a proof for the Riemann hypothesis over a period of approximately six weeks.
If verified, the discovery would represent a landmark achievement in the application of artificial intelligence to complex mathematics. However, the claim arrives amidst ongoing tension within the academic community regarding the dominance of well-funded private corporations in theoretical research.
Reflecting on the collaboration, Alpöge described the AI’s performance in the final stages of the work as exceptional, stating that Claude “flew, and danced, and sparkled” as they approached the finish line.
To support his claims, Alpöge provided computer code written in Lean, a formal verification language, which he maintains validates the findings. The global mathematical community has since begun reviewing the proof for accuracy.
The announcement coincided with Anthropic’s recent debut on the Nasdaq stock market. Alpöge acknowledged the awkward timing of the release, noting that he would have preferred to finalize the work sooner.
First proposed by German mathematician Bernhard Riemann in 1859, the Riemann hypothesis concerns the distribution of prime numbers—integers greater than one divisible only by themselves and one. Riemann’s work sought to understand the frequency and pattern of these numbers along the number line.
Due to its complexity and importance, the hypothesis is one of the seven Millennium Prize Problems established by the Clay Mathematics Institute in 2000, which offers a $1 million reward for its resolution.
While some embrace AI, others remain skeptical. Alessandro Masullo, an independent Italian mathematician who recently released an incomplete proof after 12 years of manual work, stated he avoided AI tools due to a lack of trust in their reliability.
Broader concerns have also been raised regarding the “brute force” nature of AI research. In September, OpenAI claimed a non-public version of its AI helped address another Millennium prize problem, though critics noted the immense resources used—including 10,000 AI systems—to achieve the result.
James Robinson, a mathematician at the University of Warwick, criticized this trend as “immature playground boasting” by big tech companies, arguing that the massive energy and financial costs associated with such demonstrations are problematic in the face of the global climate crisis.
Also Read
- Mexican Authorities Apprehend ‘El Chapo Isidro’ in Sinaloa, Renewing Questions Over Cartel Violence
- Pentagon Rejects Iran’s Claim of Missile Strike on US F-16 Over the Gulf
- Wes Streeting calls Putin a ‘cold-blooded killer’ as he meets Zelensky in Kyiv
- FBI Investigates Suspected Intrusion Into National Crime Information Center Database


