Anthropic researcher claims AI helped solve Riemann hypothesis
A mathematician at Anthropic claims to have used Claude to solve one of the most famous unsolved problems in mathematics, though the proof remains unverified.

The short version
- The Riemann hypothesis is one of the seven Millennium Prize Problems, which carry a $1 million prize for a verified solution.
- Alpöge used the AI chatbot Claude to formulate the proof, which he published in the Lean programming language.
- The claim comes shortly after OpenAI announced an AI-assisted solution to the Navier-Stokes existence and smoothness problem.
- The proof has not been confirmed by the Clay Mathematics Institute, Anthropic, or any other credible third party.
Why it matters
If verified, this would be a landmark moment for AI's ability to perform high-level scientific reasoning. It would validate the claim that large language models can expand the frontier of knowledge in fields like mathematics.
A claim of mathematical history
Levent Alpöge, a Turkish-American mathematician currently working at Anthropic, has made a bold claim regarding one of the most difficult puzzles in history. He asserts that he has solved the Riemann hypothesis, a problem first formulated in 1859 by Bernhard Riemann. The hypothesis focuses on how prime numbers are spread out and is widely regarded as the most formidable unresolved mystery in the field of mathematics.
Because of its difficulty, the Riemann hypothesis is categorized as one of the seven Millennium Prize Problems designated by the Clay Mathematics Institute. Each of these problems carries a $1 million prize for a solution that meets the institute's specific requirements. Many researchers refer to it as the great white whale of mathematical research due to its long-standing status as an unsolved mystery.
Alpöge published his work in the Lean programming language, which is a tool used by mathematicians to write formal proofs. However, as of now, the proof has not been confirmed by the Clay Mathematics Institute, Anthropic, or any other reputable organization. The mathematical community is still waiting to see if the logic holds up under rigorous peer review.
The role of Claude in the proof
Alpöge states that he used Anthropic's chatbot, Claude, to help formulate the proof for this complex hypothesis. He describes the experience as a collaborative effort where the AI significantly accelerated the process. He noted that while he was the primary driver, the AI eventually took the lead during the final stages of the work.
In a statement on his website, Alpöge described the experience as feeling like a divine revelation. He wrote that Claude far outpaced him near the finish line, saying the AI flew, danced, and sparkled during the final steps. This suggests that the AI was able to handle complex logical leaps that were difficult for the human researcher to navigate alone.
This interaction highlights a growing trend of humans and AI working together to tackle thorniest scientific mysteries. It follows a similar path to OpenAI's recent announcement regarding the Navier-Stokes problem. That feat was achieved using around 10,000 AI agents to solve a problem that had eluded mathematicians for decades.
Scientific discovery and AI
The potential verification of this proof would be a major milestone for the AI industry. For years, companies have spent massive amounts of money on models with the promise that they would eventually advance scientific discovery. A solved Riemann hypothesis would provide a concrete example of AI expanding the frontier of knowledge in a way humans have not been able to do alone.
OpenAI CEO Sam Altman recently echoed this sentiment during an interview with Fortune. He suggested that we are entering a golden age of scientific discovery because current models are clearly capable of expanding the frontier of knowledge. He believes this same technology will eventually work for discovering many other types of science beyond just mathematics.
For Anthropic, a successful verification would be a significant win for the company's reputation. It comes at a time when the company is trying to win over investors following its record-breaking IPO. While the timing of the announcement was close to the stock market debut, Alpöge insists that the timing was not intended to be cynical or to drive up share prices.
Context of previous AI claims
The atmosphere surrounding AI-assisted math is already tense due to previous controversies. Recently, Alpöge and another mathematician, Tristan Buckmaster, claimed that their work on a related problem was used by OpenAI without proper credit. This controversy was fueled in part by Alpöge’s current affiliation with Anthropic, the company that competes with OpenAI.
The Navier-Stokes solution announced by OpenAI in September also faced scrutiny. While it was hailed by some as a glimpse into a new era of human-AI collaboration, others viewed it as a warning of a potential jobs apocalypse. These events have set the stage for a high-stakes environment where every new AI-assisted claim is met with both excitement and skepticism.
Moving forward, the mathematical community will need to conduct a thorough review of Alpöge's work in Lean. Because the proof is currently unverified, the focus remains on whether the logic is sound or if the AI produced a hallucination. If it is correct, it will be one of the most significant mathematical achievements of the 21st century.
What to watch
- Whether the Clay Mathematics Institute officially verifies Alpöge's proof.
- How the broader mathematical community reacts to the use of Lean and Claude for high-level proofs.
- Future announcements from Anthropic regarding their model's capabilities in scientific research.
Try it
You can look up the 'Millennium Prize Problems' on the Clay Mathematics Institute website to see the other six unsolved puzzles that carry a $1 million prize.
Sources and links
- Videoyoutube.com/watch youtube.com
- Reportinggizmodo.com: the article we sourced
Read the original at gizmodo.com
First reported by gizmodo.com. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.