Prime gap
≈ 9 min read
Names in other languages
No names in other languages are held for this entry yet.
Article bodies are English and are not machine-translated. A name in another language is not a translation of the article.
A prime gap is the difference between two successive prime numbers. The n -th prime gap, denoted g_n or g(p_n) is the difference between the (n+1) th and the n -th prime numbers, i.e., : g_n = p_{n+1} - p_n For example, since the first few primes are 2, 3, 5, 7, 11..., we have g_1 = 1 , g_2 = g_3 = 2 , g_4 = 4 . The sequence g_n of prime gaps has been extensively studied; however, many questions and conjectures remain unanswered. The first 60 prime gaps are: : 1, 2, 2, 4, 2, 4, 2, 4, 6, 2, 6, 4, 2, 4, 6, 6, 2, 6, 4, 2, 6, 4, 6, 8, 4, 2, 4, 2, 4, 14, 4, 6, 2, 10, 2, 6, 6, 4, 6, 6, 2, 10, 2, 4, 2, 12, 12, 4, 2, 4, 6, 2, 10, 6, 6, 6, 2, 6, 4, 2, ... . By the definition of g_n every prime can be written as : p_{n+1} = 2 + \sum_{i=1}^n g_i.
A prime gap is the difference between two successive prime numbers. The n -th prime gap, denoted g_n or g(p_n) is the difference between the (n+1) th and the n -th prime numbers, i.e., g_n = p_{n+1} - p_n For example, since the first few primes are 2, 3, 5, 7, 11..., we have g_1 = 1 , g_2 = g_3 = 2 , g_4 = 4 . The sequence g_n of prime gaps has been extensively studied; however, many questions and conjectures remain unanswered. The first 60 prime gaps are: 1, 2, 2, 4, 2, 4, 2, 4, 6, 2, 6, 4, 2, 4, 6, 6, 2, 6, 4, 2, 6, 4, 6, 8, 4, 2, 4, 2, 4, 14, 4, 6, 2, 10, 2, 6, 6, 4, 6, 6, 2, 10, 2, 4, 2, 12, 12, 4, 2, 4, 6, 2, 10, 6, 6, 6, 2, 6, 4, 2, ... . By the definition of g_n every prime can be written as p_{n+1} = 2 + \sum_{i=1}^n g_i. In the opposite direction, the twin prime conjecture posits that g_n=2 for infinitely many integers n . |-style="vertical-align: top;" | \lim_{n\to\infty}\frac{g_n}{p_n}=0. In other words (by definition of a limit), for every \varepsilon > 0 , there is a number N such that for all n > N , g_n Hoheisel (1930) was the first to show a sublinear dependence; that there exists a constant \theta such that \pi(x + x^\theta) - \pi(x) \sim \frac{x^\theta}{\log x} \text{ as } x \to \infty, hence showing that for sufficiently large n , g_n Hoheisel obtained the possible value 32999/33000 for \theta . This was improved to 249/250 by Heilbronn, and to \theta = 3/4 + \varepsilon , for any \varepsilon > 0 , by Chudakov. A major improvement is due to Ingham, who showed that for some positive constant c, if \zeta(1/2 + it) = O(t^c) then \pi(x + x^\theta) - \pi(x) \sim \frac{x^\theta}{\log x} for any \theta > (1 + 4c)/(2 + 4c). Here, O refers to the big O notation, ζ denotes the Riemann zeta function and π the prime-counting function. Knowing that any is admissible, one obtains that θ may be any number greater than 5/8. Since 5/8+ε n^{2/3} , it follows that there is always a prime number between n 3 and , if n is sufficiently large. In 2016, Dudek gave an explicit version of Ingham's result: there are primes between consecutive cubes for all n > e^{e^{33.217}}\approx1.01\cdot10^{115809466034000} . The Lindelöf hypothesis would imply that Ingham's formula holds for c any positive number: but even this would not be enough to imply that there is a prime number between n 2 and for n sufficiently large (see Legendre's conjecture). To verify this, a stronger result such as Cramér's conjecture would be needed. Huxley in 1972 showed that one may choose . A result, due to Baker, Harman and Pintz in 2001, shows that θ may be taken to be 0.525. The above describes limits on all gaps; another area of interest is the minimum gap size. The twin prime conjecture asserts that there are always more gaps of size 2, but remains unproven. In 2005, Daniel Goldston, János Pintz and Cem Yıldırım proved that \liminf_{n\to\infty}\frac{g_n}{\log p_n} = 0 and 2 years later improved this to \liminf_{n\to\infty}\frac{g_n}{\sqrt{\log p_n}(\log\log p_n)^2} In 2013, Yitang Zhang proved that \liminf_{n\to\infty} g_n meaning that there are infinitely many gaps that do not exceed 70 million. A Polymath Project collaborative effort to optimize Zhang's bound managed to lower the bound to 4680 on July 20, 2013. In November 2013, James Maynard introduced a new refinement of the GPY sieve, allowing him to reduce the bound to 600 and also show that the gaps between primes m apart are bounded for all m. That is, for any m there exists a bound Δ m such that for infinitely many n. Using Maynard's ideas, the Polymath project improved the bound to 246. On August 28, 2026, an argument discovered by GPT-5.6 Sol and presented by Hanxin Zhang claimed an improvement on the Polymath8b bound from 246 to 240. 3 days later Julia Stadlmann submitted a preprint developing new techniques and also claiming to reduce the bound to 240. In work dating 1 September, 2026 Shiva Kintali with unspecified AI agents claimed an improvement of the bound to 236. On September 3 Axiom Math with AxiomProver claimed an improvement of the bound to 212 and a few hours before (on September 2 in USA) OpenAI using GPT-6 Astra produced a further claim to the bound to 186. As of September 2026, none of those claims has been peer-reviewed, leaving the established bound at 246. Assuming the Elliott–Halberstam conjecture and its generalized form, the bound has been reduced to 12 and 6, respectively. \limsup_{n\to\infty}\frac{g_n}{\log p_n}=\infty. In 1938, Robert Rankin proved the existence of a constant such that the inequality g_n > \frac{c\ \log n\ \log\log n\ \log\log\log\log n}{(\log\log\log n)^2} holds for infinitely many values of n, improving the results of Westzynthius and Paul Erdős. He later showed that one can take any constant , where γ is the Euler–Mascheroni constant. The value of the constant c was improved in 1997 to any value less than 2e γ . Paul Erdős offered a $10,000 prize for a proof or disproof that the constant c in the above inequality may be taken arbitrarily large. This was proved to be correct in 2014 by Ford–Green–Konyagin–Tao and, independently, James Maynard (the two works were submitted to arXiv within one day of each other). The result was further improved to g_n > \frac{c\ \log n\ \log\log n\ \log\log\log\log n}{\log\log\log n} for infinitely many values of n by Ford–Green–Konyagin–Maynard–Tao. In the spirit of Erdős' original prize, Terence Tao offered US$10,000 for a proof that c may be taken arbitrarily large in this inequality. In 2026, a proof and formalization in Lean claiming authorship by AI model GPT-5.6 Sol was uploaded to Github, positing an improvement on the bound to: g_n > \frac{c\ \log n\ \log\log n}{\log\log\log\log n} The claimed proof has not yet been fully verified. Later that year, OpenAI released a proof and formalization in Lean improving the bound to: g_n > \frac{c\ \log n\ (\log\log n)^2\ \log\log\log\log n}{(\log\log\log n)^2} The claimed proof has not yet been fully verified. Lower bounds for chains of primes have also been determined. g_n = O(\sqrt{p_n} \log p_n), using the big O notation. (In fact this result needs only the weaker Lindelöf hypothesis, if one can tolerate an exponent infinitesimally larger than 1/2. ) Dudek also proved an explicit version of Cramer's result (also assuming Riemann hypothesis) that is, for all n ≥ 2, g_n As noted by Dudek, for sufficiently large n one may reduce the constant 4/\pi to be any constant greater than 1. In the same article, Cramér conjectured that the gaps are far smaller. Roughly speaking, Cramér's conjecture states that g_n = O\!\left((\log p_n)^2\right)\!, a polylogarithmic growth rate slower than any exponent . Cramér's model, under which he made the conjecture, was oversimplified (assuming some events are statistically independent when they are dependent) and thus not very accurate (see Cramér's conjecture), but after further investigations, new heuristics were found which became strong evidence that the conjecture is true. As this matches the observed growth rate of prime gaps, there are a number of similar conjectures. Firoozbakht's conjecture is slightly stronger, stating that p n 1/n is a strictly decreasing function of n, i.e., for all n ≥ 1, (p_{n+1})^{1/(n+1)} If this conjecture is true, then . It implies a strong form of Cramér's conjecture but is inconsistent with the heuristics of Granville and Pintz, which suggest that infinitely often for any , where γ denotes the Euler–Mascheroni constant. Polignac's conjecture states that every positive even number k occurs as a prime gap infinitely often. The case k = 2 is the twin prime conjecture.
Text derived from https://en.wikipedia.org/w/index.php?oldid=1378725169 under CC BY-SA 4.0; modifications have been made.