QUICK TAKE | Has Claude Helped Solve a 167-Year-Old Maths Mystery? Seven Things to Know

1. What is the Riemann hypothesis?
Proposed by Bernhard Riemann in 1859, it concerns the distribution of prime numbers: numbers such as 2, 3, 5 and 7, divisible only by themselves and one. It predicts how far the count of primes can deviate from its expected average as numbers grow larger.
2. What exactly does it say?
The hypothesis concerns a mathematical object called the Riemann zeta function. It says all its “non-trivial” zeros lie on a particular line, where the real part equals one-half. Establishing this for every such zero would provide extraordinarily precise control over estimates involving prime numbers.
3. Who is claiming to have proved it?
Levent Alpöge, a mathematician working at Anthropic, announced the proposed breakthrough on his personal website. He credits collaboration with Claude and claims the method also resolves the broader Generalized Riemann Hypothesis. He describes a human-guided collaboration in which Claude supplied crucial mathematical ideas.
4. What has he actually published?
His four-page announcement links to files containing the proposed formal proof in Lean, a language used for computer-checkable mathematics. Alpöge says an attempt to produce a detailed written explanation exceeded 250,000 pages. He plans to develop a clearer account and simplify the argument.
The Wealth Issue 2026
06 Nov 2026 - Vol 05 | Issue 45
Portraits of Dreamers and Doers under 35
5. Does a computer-checkable proof settle the claim?
Calling an argument formalised does not itself establish independent verification. Reviewers must check that the files work, that the formal statement matches the intended problem and that the proof rests on acceptable foundations. Alpöge’s assertion that the arguments contain no assumptions is part of the claim being evaluated.
6. Has the million-dollar prize been won?
There is no verified award announcement in the material reviewed. Clay Mathematics Institute still lists the problem as unsolved. Its prize rules require publication in a qualifying outlet, at least two years after publication and general acceptance by the global mathematics community before consideration.
7. Would success mean AI can solve every mathematical problem?
A verified proof would be a landmark for mathematics and AI-assisted research. It would not establish that every unsolved problem is within AI’s reach. The immediate question is whether this particular argument proves what Alpöge claims. The supplied report’s declaration that all frontiers can be breached goes beyond the evidence.
