Science & Environment
Anthropic scientist says AI helped solve math's famous Riemann puzzle
Reported by Vivek Joshi (Copy Editor) · The Hindu - Science ·

Mathematician Levent Alpöge from the U.S. artificial intelligence company Anthropic stated on Friday, November 20, 2026, that he found a proof for the Riemann hypothesis using the company model Claude. The problem is one of the most important unsolved puzzles in mathematics.
Details of the Proof
Named after German mathematician Georg Bernhard Riemann who proposed it in 1859, the hypothesis deals with the spread of prime numbers, which form all whole numbers. Experts have spent over 160 years trying to solve it. This puzzle is also one of seven problems carrying a one-million-dollar reward from the Clay Mathematics Institute. Alpöge shared over 1.7 crore lines of computer code containing his claimed proof and stated he also proved a stronger result known as the generalised Riemann hypothesis. He did not write a traditional paper and noted he does not yet fully understand every argument the AI used.
Verification and Checks
Independent mathematicians, including Andrew Sutherland from the Massachusetts Institute of Technology, checked the published code and confirmed it passes several tough tests. Two separate proof-checking systems also accepted the result.
Context
Solving the Riemann hypothesis is a major milestone because prime numbers are foundational to many areas of modern mathematics and digital security systems. When researchers make progress on problems of this scale, it often pushes forward new methods in computing and advanced science.
ExplainerWhy this matters
What Happened
A researcher at Anthropic used an AI model named Claude to generate computer code that purportedly proves the Riemann hypothesis, a famous 165-year-old math problem about prime numbers.
Why It Matters
Proving this hypothesis could solve a century-old mystery in mathematics and unlock a one-million-dollar prize. More importantly, it raises new questions about how scientists use artificial intelligence to solve complex problems even when humans do not fully understand every step the AI takes.
What Happens Next
Other mathematicians will now spend a lot of time carefully studying the massive block of code and testing all the logic to see if the proof holds up under strict academic review.
More from today
Science & Environment• UpdatedExplainerSource: The HinduWorth knowing
Anthropic AI and Researcher Release Code to Prove Riemann Hypothesis.
Reported by Imran Qureshi (Staff Writer) · The Hindu - Science ·
