Riemann Hypothesis Solved? What’s Verified in 2026
Quick answer: Yes, a formally verified proof now exists. On November 20, 2026, Levent Alpöge released a proof of the Riemann Hypothesis developed with the AI model Claude, written in the Lean proof assistant. On November 21, the Lean and Mathlib communities issued a joint statement certifying that “both the statement and the proof” are correct. A readable human write-up and any decision on the $1 million Clay prize are separate matters still to come.
Status checked November 21, 2026 (U.S. Central time). Verification status is reported from the dated primary sources linked beside each claim.
Prime numbers get rarer as numbers grow, yet their exact positions are hard to predict. The Riemann Hypothesis describes how closely the count of primes follows a smooth mathematical estimate, and it does so through the zeros of a function called the Riemann zeta function. For 167 years nobody could prove it. This guide starts with the news, what has been verified and by whom, and then explains the mathematics in plain English.
Jump to: Proof status · Plain-English explanation · Claude and Lean
Has the Riemann Hypothesis Been Proven?
As of November 21, 2026, the answer from the formal-verification community is yes. Here is what each source reports:
| Question | Status | Source |
|---|---|---|
| Was a proof released? | Yes. Levent Alpöge announced it on November 20, 2026, as a Lean library of about 17 million lines (175,187 files in the original release), developed with Claude. | Alpöge’s announcement note |
| Does the Lean theorem state the real Riemann Hypothesis? | Yes. The statement was compared against Mathlib’s existing definition with comparison tools, and the joint statement certifies it. | Lean and Mathlib joint statement |
| Has the formal proof been checked? | Yes. Every file compiles, the headline theorems use only Lean’s three standard axioms, and several independent proof checkers accept it. A few extra long-running checks were still finishing. | Andrew Sutherland’s verification report (MIT) |
| Is there a readable mathematical write-up? | Not yet. The announcement presents the Lean code; Alpöge says a proper explanatory write-up will follow. | Announcement note |
| Has Clay recognized it or awarded the prize? | No. Clay’s page still labels RH unsolved, and its rules require a waiting period after publication. | Clay Mathematics Institute |
In the joint statement, the Lean and Mathlib communities, endorsed by the Lean FRO and the Mathlib Initiative, write: “We hereby certify that both the statement and the proof of the Riemann Hypothesis are correct.” The verified result also covers the Generalized Riemann Hypothesis for Dirichlet L-functions, a family of functions closely related to zeta. It does not cover every “generalized” version mathematicians study.
What Is the Riemann Hypothesis in Simple Terms?
Start with the zeta function. For a complex number \(s\) whose real part is greater than 1, it is the infinite sum
\[ \zeta(s) = \sum_{n=1}^{\infty} \frac{1}{n^s} = 1 + \frac{1}{2^s} + \frac{1}{3^s} + \cdots \]
For example, \(\zeta(2) = 1 + \tfrac14 + \tfrac19 + \cdots = \pi^2/6\), a famous result of Leonhard Euler. The sum stops making sense when the real part of \(s\) is 1 or less, but a method called analytic continuation extends the function to the whole complex plane, except for one point at \(s = 1\). Bernhard Riemann did this in 1859, and the zeros that matter belong to this extended function.
The zeros come in two kinds:
- Trivial zeros at \(-2, -4, -6, \ldots\). These are well understood.
- Nontrivial zeros, all known to lie in the critical strip, where the real part is between 0 and 1.
The Riemann Hypothesis states that every nontrivial zero has real part exactly \(\tfrac12\). In other words, they all sit on one vertical line, the critical line. The first three nontrivial zeros with positive imaginary part are approximately
\[ \tfrac12 + 14.1347i, \qquad \tfrac12 + 21.0220i, \qquad \tfrac12 + 25.0109i. \]
Why Does the Zeta Function Tell Us About Primes?
The bridge is Euler’s product formula, which rewrites the sum over all whole numbers as a product over the prime numbers:
\[ \zeta(s) = \prod_{p\ \text{prime}} \frac{1}{1 – p^{-s}} = \frac{1}{1-2^{-s}}\cdot\frac{1}{1-3^{-s}}\cdot\frac{1}{1-5^{-s}}\cdots \qquad (\operatorname{Re}(s) > 1) \]
Each prime contributes one factor. This works because every whole number factors into primes in exactly one way. So information about zeta is secretly information about primes.
Now count primes. Let \(\pi(x)\) be the number of primes up to \(x\). The Prime Number Theorem (1896) says \(\pi(x)\) is close to \(x/\ln x\), where \(\ln\) is the natural logarithm, and even closer to the logarithmic integral \(\operatorname{li}(x)\):
| \(x\) | Actual \(\pi(x)\) | \(x/\ln x\) | \(\operatorname{li}(x)\) | Relative error of \(\operatorname{li}(x)\) |
|---|---|---|---|---|
| 1,000 | 168 | 145 | 178 | 5.72% |
| 1,000,000 | 78,498 | 72,382 | 78,628 | 0.165% |
| 1,000,000,000 | 50,847,534 | 48,254,942 | 50,849,235 | 0.0033% |
| 1,000,000,000,000 | 37,607,912,018 | 36,191,206,825 | 37,607,950,281 | 0.000102% |
Here \(\operatorname{li}(x)\) is the integral of \(1/\ln t\) from 0 to \(x\). Estimates are rounded to whole numbers; the relative error \(100\,|\operatorname{li}(x) – \pi(x)|/\pi(x)\) uses unrounded values.
Notice the last row: at one trillion, \(\operatorname{li}(x)\) is off by about 38,263 primes, yet that is only about 0.0001 percent. Absolute and percentage errors tell different stories.
Riemann found an exact formula in which each nontrivial zero adds a wave-like correction to the smooth estimate. The further a zero lies from the critical line, the larger its correction can grow. The Riemann Hypothesis is equivalent to a strong bound on the error:
\[ \pi(x) = \operatorname{li}(x) + O\!\left(\sqrt{x}\,\ln x\right). \]
The \(O\) means the error stays below a constant times \(\sqrt{x}\,\ln x\) for all large \(x\). It does not mean primes are evenly spaced; their gaps still vary. It means the running total never drifts far from the estimate.
Checking Zeros Versus Checking a Proof
Computers had already confirmed that more than 12.3 trillion zeros lie on the critical line (Platt and Trudgian, 2021). That is strong evidence, but each check covers one case, and infinitely many remain. A proof is a general argument that covers every zero at once. When Lean checks a proof, it checks that general argument step by step, which is a completely different kind of evidence from testing examples.
What Did Claude Do, and What Did Lean Check?
| Name | Role |
|---|---|
| Levent Alpöge | Mathematician at Anthropic who announced the result. |
| Claude | Anthropic’s AI model; according to Alpöge’s account, his collaborator in developing the argument and the formalization. |
| Lean | The proof assistant in which the proof is written and mechanically checked. |
| Mathlib | Lean’s shared library of formalized mathematics, which supplies the standard definition of the Riemann Hypothesis. |
| Independent verifiers | People such as Andrew Sutherland at MIT, plus the Lean and Mathlib communities, who checked the proof and that its conclusion matches the intended statement. |
How the work was divided between Alpöge and Claude has not been described in detail, and we do not guess at it here.
RH, Quasi-RH and GRH: What Is the Difference?
- Quasi-Riemann Hypothesis (October 2026): an OpenAI model proved that zeros stay inside a narrower strip, for example with real parts between 1/12 and 11/12. That is weaker than RH. See our quasi-Riemann Hypothesis explainer.
- Riemann Hypothesis (November 2026): every nontrivial zero lies exactly on the line at real part 1/2.
- Generalized Riemann Hypothesis for Dirichlet L-functions: the same statement for a family of related functions used to study primes in arithmetic progressions. It is included in the verified release.
What Changes Now That RH Is Proved?
Hundreds of published theorems begin with “Assume the Riemann Hypothesis.” With a verified proof, those results become unconditional, including sharp bounds on gaps between primes and on how primes split among arithmetic progressions. Expect mathematicians to spend years searching the 17 million lines for the key ideas and shorter arguments.
Two things do not change. Proving RH does not by itself provide a fast algorithm for factoring the large numbers used in RSA encryption, so your online banking is not suddenly at risk. And the Clay prize follows its own process: Clay requires publication in a Qualifying Outlet, at least two years since publication, and general acceptance in the global mathematics community before it considers a proposed solution.
A Short History of the Riemann Hypothesis
| Year | Event |
|---|---|
| 1859 | Riemann states the hypothesis in an eight-page paper on counting primes. |
| 1896 | Hadamard and de la Vallée Poussin prove the Prime Number Theorem. |
| 1900 | Hilbert includes it as problem 8 in his famous list of 23 problems. |
| 1914 | Hardy proves that infinitely many zeros lie on the critical line. |
| 2000 | The Clay Institute names it a $1 million Millennium Prize Problem. |
| 2021 | Platt and Trudgian confirm more than 12.3 trillion zeros lie on the line. |
| Oct 2026 | An AI-generated proof of the weaker quasi-Riemann Hypothesis is released. |
| Nov 20, 2026 | Alpöge releases a Lean proof of RH developed with Claude. |
| Nov 21, 2026 | The Lean and Mathlib communities certify the statement and the proof. |
Common Misconceptions
“Checking trillions of zeros already proved it.” No. Examples can never settle a statement about infinitely many zeros. Only a proof can.
“RH says primes are evenly spaced.” No. Prime gaps still vary. RH bounds how far the total count of primes can stray from \(\operatorname{li}(x)\).
“A Lean proof means humans understand why RH is true.” Not yet. Lean certifies correctness; understanding the ideas is the work that comes next.
Watch: The Riemann Hypothesis in Video
If you learn better by watching, these videos are worth your time.
Check Your Understanding
Five quick questions to see how much of this has stuck. Pick an answer and press Submit; you will see a short explanation either way.
Answer each question and press Submit to see your result right away.
Frequently Asked Questions
Has the Riemann Hypothesis been proven?
Yes, according to the formal-verification community. Levent Alpöge released a Lean proof developed with Claude on November 20, 2026, and on November 21 the Lean and Mathlib communities certified that both the statement and the proof are correct. A readable write-up and any Clay prize decision are still to come.
Where can I read the Riemann Hypothesis proof?
Start with Alpöge’s four-page announcement note at alpo.ge, which links the Lean files. The Lean and Mathlib joint statement explains how the statement and proof were checked, and Andrew Sutherland’s MIT page lists each verification run. A traditional readable paper has not been released yet.
Why does Clay still list the Riemann Hypothesis as unsolved?
Clay only considers a solution after publication in a Qualifying Outlet, at least two years since publication, and general acceptance by the global mathematics community. Its listing reflects that process, not a judgment that the formal proof is wrong.
Who gets the 1 million dollar prize?
Nobody yet. Clay’s scientific advisory board will decide eligibility after its waiting period, and how to credit work developed with an AI system has not been settled publicly.
Does proving the Riemann Hypothesis break encryption?
No. RSA encryption depends on how hard it is to factor large numbers, and proving RH does not by itself give a fast factoring method.
Keep Exploring
- Quasi-Riemann Hypothesis Explained: OpenAI’s Proof
- Twin Primes (Math Dictionary)
- 10 Famous Math Problems and the History Behind Them
Sources and Further Reading
- Lean and Mathlib communities: Lean verification of the Riemann Hypothesis (Nov 21, 2026)
- Alpöge, “Remark” (announcement note, Nov 20, 2026)
- Andrew Sutherland, MIT: RH Lean proof verification files
- Clay Mathematics Institute: The Riemann Hypothesis
- Clay Mathematics Institute: Rules for the Millennium Prizes
- NIST Digital Library of Mathematical Functions: Riemann zeta function
- Platt and Trudgian, The Riemann hypothesis is true up to 3·10^12 (arXiv)
Related to This Article
More math articles
- Sustainable Development and Population-Urban-Environment Synthesis
- Federalist No. 10 and factions
- 8th Grade OST Math Worksheets: FREE & Printable
- How to Find 3D Shapes Nets?
- 7th Grade Math Worksheets for Arizona (AASA): 95 Free Printable PDFs
- Due Process and Privacy: Rights, Procedures and Limits
- Breathing and Gas Exchange
- FREE 8th Grade PARCC Math Practice Test
- Theories and Levels of Analysis
- ACT Math Formulas


What people say about "Riemann Hypothesis Solved? What’s Verified in 2026 - Effortless Math"?
No one replied yet.