Twin Primes Thin Out Faster Than Primes Do
Published 10/26/2026 · 3 min read · Everyday calculators
Lena Hoffmann — Science & education writer at OneKitly
Mathematics · Physics
Checked against 2 sources
Twin primes are pairs two apart: 3 and 5, 11 and 13, 41 and 43. Below a hundred there are eight such pairs against twenty-five primes; below a thousand, thirty-five pairs against a hundred and sixty-eight primes; below a hundred thousand, 1,224 pairs against 9,592 primes. Primes themselves get rarer as numbers grow, and twins get rarer faster — the pairs fall from about a third of the prime count to about an eighth over that range. Whether the pairs eventually stop is the twin prime conjecture, and it is open. What is known is that pairs no more than a few hundred apart certainly do go on for ever, a result proved in 2013 and sharpened rapidly afterwards.
Below 100 there are 25 primes and 8 twin pairs. Below 100,000 there are 9,592 primes and 1,224 pairs — the share drops from 32 % to 13 %.
Rare enough that their reciprocals add up to a finite number
Adding up one over every prime — a half, a third, a fifth, a seventh and so on for ever — grows without bound, slowly but definitely, and that divergence is one of the classical proofs that primes never run out. Do the same over twin primes only and the sum converges to a finite value, a result from 1919. The consequence is unwelcome: the argument that works for primes cannot work for twins, because the very fact that makes it work — the sum running away — is false here. Any proof that twins are infinite has to come from somewhere else, and that is a large part of why the conjecture is still open.
Three apart is impossible, and that is the whole reason two is special
Two primes cannot differ by an odd number unless one of them is two, because two odd numbers always differ by an even amount. Two is therefore the smallest gap that can occur infinitely often, which is exactly why the conjecture is stated about it. Gaps of four — cousin primes — and six — sexy primes — have their own conjectures with the same status, and all of them are special cases of a general pattern that predicts how often any given gap should appear. That general prediction matches the counts extremely well, which is more evidence and, once again, not a proof.
| Up to | Primes | Twin pairs | Pairs per prime |
|---|---|---|---|
| 100 | 25 | 8 | 32.0 % |
| 1,000 | 168 | 35 | 20.8 % |
| 10,000 | 1,229 | 205 | 16.7 % |
| 100,000 | 9,592 | 1,224 | 12.8 % |
Frequently asked questions
- What was actually proved in 2013?
- That some fixed gap occurs infinitely often — not necessarily two. The first proof established a bound in the tens of millions, meaning there is at least one even number below that size which appears as a prime gap for ever, without identifying which. A collaborative effort pushed the bound down to a few hundred within a year. Getting from a few hundred to two would settle the conjecture, and the known techniques appear to have a hard floor somewhere above two — the last step is not a matter of more effort along the same line.
- Are there twin primes among very large numbers?
- Yes, and record hunters find larger ones regularly — the current records have hundreds of thousands of digits and are found by searching specially shaped candidates that are cheap to test. Those discoveries prove nothing about the conjecture, since finding a large example is compatible with the supply eventually stopping. They do keep confirming that twins remain plentiful far beyond any range a person could examine, which is consistent with everything the general prediction says and, as with every other line of evidence here, not the thing anybody actually wants.
Articles you may find interesting
All guides →Related tools
Sources
Spotted a mistake in this article?