# What is the Riemann hypothesis?

Japanese: リーマン予想 (りーまんよそう)

**In one line:** The Riemann hypothesis says that all the non-trivial zeros of the zeta function lie on one straight line (real part 1/2). It is the key to how prime numbers are spread out.

Many people think the Riemann hypothesis is just a very hard sum. It is really a guess about prime numbers. Primes like 2, 3, 5 and 7 seem to pop up at random. In 1859, a mathematician named Riemann had an idea. A formula called the zeta function hits zero at special points. If all of them sit on one straight line, we can almost read where primes appear. The first ten trillion points are on the line. Nobody has shown that all of them are. There is a million-dollar prize. In November 2026, someone said they proved it with an AI. People are checking it now. In math, the long part is not “I solved it!” It is “Yes, you really did.”

**Common mix-up:** Not yet. In November 2026 someone announced a proof, but until other mathematicians and computers finish checking it, the problem is officially still open.

**Primes and the zeta function**

A prime is a number you can only divide by 1 and itself. Primes look scattered, but we know roughly how often they appear among big numbers (the prime number theorem). The Riemann hypothesis says how small the errors around that average are. In his 1859 paper, Riemann tied the zeros of the zeta function to counting primes, and guessed that every interesting zero has real part 1/2.

**Why is there a prize?**

The problem was on David Hilbert's list of 23 problems in Paris in 1900, and on the seven Millennium Prize Problems the Clay Mathematics Institute announced on May 24, 2000. Each prize is one million dollars. Of the seven, only the Poincaré conjecture, proved by Grigori Perelman in preprints in 2002 and 2003, has been accepted as solved.

**The November 2026 announcement**

Levent Alpöge of Anthropic says that, working with Claude, he believes he found a proof, and that the method also covers the Generalized Riemann Hypothesis. There is no normal paper yet. The proof was posted in Lean, a proof-checking language. News reports say a human-readable draft would run past 250,000 pages. Whether it is correct is still being checked.

**How people use it**

- “If the Riemann hypothesis really fell, the textbooks will change.”
- “The proof is out. Now who is going to read all of it?”

**FAQ**

- Has the Riemann hypothesis been proved? — Not yet. In November 2026 someone announced a proof, but until other mathematicians and computers finish checking it, the problem is officially still open.
- What would we learn if it is true? — We could estimate how primes are spaced out almost as well as possible. Many results that start with “if the Riemann hypothesis is true” would then hold with no “if.”
- How big is the prize? — It is one of the seven Millennium Prize Problems chosen by the Clay Mathematics Institute in 2000. The prize is one million dollars, given only after mathematicians accept the proof.
- What was announced in November 2026? — Levent Alpöge, a mathematician at Anthropic, wrote on his website that he believes he found a proof while working with the AI Claude. The proof is written in Lean, a language computers can check, and checking is under way.

Topic: Science
Published: 2026-11-21T07:30:00+09:00
Source URL: https://eli5.net/en/w/riemann-hypothesis
日本語: https://eli5.net/w/riemann-hypothesis
