Goldbach’s conjecture
Is every even number greater than 2 the sum of two primes?
Summary
AI-assisted claims
Every even integer greater than 2 is a prime plus a product of at most 18 primes
“We prove that every even integer $N \ge 4$ can be written as $N = p + \eta$ with $p$ prime and $\eta \ge 2$ a product of at most $18$ primes, counted with multiplicity.”
- By
- Wren Aquila
- AI involvement
- Text declaration: Primarily AI-generated.
- Where
- Hexagon
- Source last read
- 14 Nov 2026
Claims as their sources state them. Listing a claim is not a check by Papex; the status stays the source’s.
Why it is open
In a letter to Leonhard Euler dated 7 June 1742, Christian Goldbach stated a conjecture about sums of primes in two forms; Euler replied on 30 June that the first follows from the statement that every even number is a sum of two primes, which Goldbach had told him earlier [2]. The strong, or binary, conjecture says that every even integer greater than 2 is the sum of two primes; the weak, or ternary, conjecture says that every odd integer greater than 5 is the sum of three primes [7]. The strong form implies the weak one: subtract 3 from the odd number and write the remainder as two primes [7]. This page is about the strong form, the one in the catalogue.
In 1900 David Hilbert mentioned Goldbach's problem in his eighth problem, as a question that a full discussion of Riemann's prime number formula might one day answer [2] (see the Papex page Riemann hypothesis). In 1912 Edmund Landau listed it among four problems about primes that he considered "unattackable at the present state of science" [2]. A resolution would be a proof for every even number greater than 2, or one even number that is not a sum of two primes. This page's status follows the Wikipedia article on Landau's problems, which says that as of 2026 all four problems are unresolved [12].
Both main tools stop short. The circle method counts representations through an integral that is split into "major arcs", near fractions with small denominators, and "minor arcs", everywhere else. Harald Helfgott explains that with three primes the major arcs dominate, but with two they are not expected to, and that because no one can prove precise estimates on the minor arcs, "the binary Goldbach conjecture is still very much out of reach" [7]. Sieve methods meet the parity problem, a heuristic principle formulated by Atle Selberg that they cannot tell apart integers with an even and an odd number of prime factors [2].
What is known
Almost all even numbers (1924–1975). Assuming the generalised Riemann hypothesis (GRH), G. H. Hardy and J. E. Littlewood showed in 1924 that the number E(X) of even numbers up to X that are not sums of two primes is at most about X^(1/2) [2]. Without that assumption, Johannes van der Corput, Theodor Estermann and Nikolai Chudakov proved independently in 1937–1938 that almost all even numbers are sums of two primes [2, 7], and Hugh Montgomery and Robert Vaughan improved this in 1975 to E(X) ≤ X^(1 − δ) for some δ > 0 [2]. Hardy and Littlewood also conjectured an asymptotic formula for the number of representations [2, 10].
A prime plus an almost prime (1920–2025). In 1920 Viggo Brun's sieve wrote even numbers as sums of two numbers with at most nine prime factors each [2]. Chen Jingrun proved (1966, 1973) that every sufficiently large even number is a prime plus a number with at most two prime factors [2, 7]. In a paper published in 2025, Matteo Bordignon, Daniel Johnston and Valeriia Starichkova proved this for every even number larger than exp(exp(32.7)) [9]; assuming GRH, Bordignon and Starichkova had obtained the threshold exp(exp(15.85)) [8]. János Pintz notes that most results for Goldbach's equation p + p′ = N carry over to p − p′ = N, which includes twin primes [2] (see the Papex page Twin prime conjecture).
Sums of more primes (1937–2014). Ivan Vinogradov proved in 1937 that every sufficiently large odd number is a sum of three primes [7]. Olivier Ramaré showed in 1995 that every even number is a sum of at most six primes, and Terence Tao that every odd number greater than 1 is a sum of at most five; Tao's proof relies on computer checks of the binary conjecture up to 4·10¹⁴ and of the Riemann hypothesis up to height 3.29·10⁹ [6, 7]. Assuming GRH, Jean-Marc Deshouillers and co-authors completed in 1997 a proof that every odd number above 5 is a sum of three primes [1, 7].
The weak conjecture (2013). The smallest threshold above which Vinogradov's theorem had been proved was about 2·10¹³⁴⁶, far beyond any computation [7]. In preprints of 2012–2013, Helfgott gave an analytic argument for every odd number from 10²⁷ upward; with David Platt he had already checked by computer all odd numbers up to 8.875·10³⁰, and his preprint "The ternary Goldbach conjecture is true" presents the two together as a proof [3, 4]. The argument also uses Platt's computer verification that the zeros of certain Dirichlet L-functions, up to a fixed height, lie on the critical line, a finite piece of GRH, and the binary check up to 4·10¹⁸ [7]. On 14 October 2026 Helfgott's publications page still listed these papers as preprints, to be superseded by a book, The ternary Goldbach problem, "to appear in Ann. of Math. Studies" [11]. The Wikipedia article used for this page's status describes the 2013 work as "a full proof of Goldbach's weak conjecture" [12]. Helfgott wrote in 2014 that the strong conjecture "remains out of reach" [7].
Computation (to 2014). Tomás Oliveira e Silva, Siegfried Herzog and Silvio Pardi confirmed the conjecture for every even number up to 4·10¹⁸; their counts of minimal Goldbach partitions, the representations that use the smallest possible prime, agree with predictions from the prime k-tuple conjecture of Hardy and Littlewood [5].
Recent claims. In June 2026 Jiamin Li and Jianya Liu posted a preprint, revised in August and not peer reviewed, claiming what they call Proposition (1 + 1.9): every sufficiently large even number N can be written as N = p + rq, where p and q are primes and r is 1 or a prime at most q^0.9 [10]. In their scale, Proposition (1 + 2) is Chen's theorem and (1 + 1) is essentially the binary conjecture; they also claim the exponent 1.4 under the Elliott–Halberstam conjecture, a conjecture on primes in arithmetic progressions, and an analogue for twin primes [10]. On 14 October 2026 the Wikipedia article still listed Goldbach's conjecture as unresolved [12].
Main approaches
The circle method. Write the number of representations as an integral of exponential sums over primes, estimate it precisely near fractions with small denominators and bound it elsewhere [7]. Strongest evidence: Vinogradov's theorem and Helfgott's ternary work [3, 4, 7], and the almost-all results for two primes [2]. Main difficulty: for two primes the minor arcs are not expected to be smaller than the major arcs, and no one knows how to estimate them precisely [7].
Sieve methods. Count the primes p for which N − p has few prime factors, as Brun, Alfréd Rényi and Chen did with successively stronger sieves [2]. Strongest evidence: Chen's theorem and its explicit forms [2, 8, 9], and the claimed Proposition (1 + 1.9) [10]. Main difficulty: the parity problem [2].
Explicit bounds and computation. Make an analytic theorem explicit, then check every case below its threshold by computer, as was done for three primes [3, 4, 7]. Strongest evidence: the ternary case, where the analytic threshold 10²⁷ fell below the checked range 8.875·10³⁰ [3, 4]. Main difficulty: no analytic theorem gives two primes for all large even numbers, and even the explicit threshold for Chen's theorem, exp(exp(32.7)), is far beyond the 4·10¹⁸ checked [5, 9].
Assuming GRH. GRH extends the Riemann hypothesis to Dirichlet L-functions [7], and several steps become possible under it: the ternary conjecture follows [1], and the count of exceptions drops to about X^(1/2) [2]. Strongest evidence: these conditional theorems, and Helfgott's use of a finite verification of GRH [1, 2, 7]. Main difficulty: GRH is unproven [7], and even under it the best bound on exceptions known in 2009 was about X^(1/2), not zero [2].
Open sub-questions
- Fewer exceptions. Can the unconditional bound E(X) ≤ X^(1 − δ) be brought close to the X^(1/2) known under GRH [2]?
- Explicit Chen. Can the threshold exp(exp(32.7)) in Chen's theorem [9], or exp(exp(15.85)) under GRH [8], be lowered towards the range already checked by computer [5]?
- Between Chen and Goldbach. If the (1 + 1.9) argument holds, how far can the exponent be lowered without the Elliott–Halberstam conjecture, under which the authors claim 1.4 [10]?
- Without zeros. Helfgott points to routes to Vinogradov's theorem that avoid zeros of L-functions [7]; can such a route be made explicit enough to give the ternary statement above a threshold that computers can reach?
- Statistics. Do minimal Goldbach partitions beyond 4·10¹⁸ keep matching the Hardy–Littlewood predictions [5]?
Work on this in Papex
If you have a testable answer to one of the sub-questions above, you can draft it as a private hypothesis in Papex, or prepare the draft from Claude or ChatGPT with the Papex connector. You review the exact version, sign it and decide whether to publish it; a public record that names this problem is listed on this page. A Papex record shows who signed which version and when. It is not peer review and does not mean the hypothesis is correct.
Key references
- Deshouillers, J.-M., Effinger, G., te Riele, H. & Zinoviev, D. (1997). A complete Vinogradov 3-primes theorem under the Riemann hypothesis. Electronic Research Announcements of the American Mathematical Society 3, 99–104. 10.1090/S1079-6762-97-00031-0 · Checked 14 Oct 2026
- Pintz, J. (2009). Landau’s problems on primes. Journal de Théorie des Nombres de Bordeaux 21(2), 357–404. 10.5802/jtnb.676 · Checked 14 Oct 2026
- Helfgott, H. A. & Platt, D. J. (2013). Numerical verification of the ternary Goldbach conjecture up to 8.875·10^30. Experimental Mathematics 22(4), 406–409. 10.1080/10586458.2013.831742 · Checked 14 Oct 2026
- Helfgott, H. A. (2013). The ternary Goldbach conjecture is true. arXiv:1312.7748 (v1 30 December 2013, v2 17 January 2014). Preprint. Source · Checked 14 Oct 2026
- Oliveira e Silva, T., Herzog, S. & Pardi, S. (2014). Empirical verification of the even Goldbach conjecture and computation of prime gaps up to 4·10^18. Mathematics of Computation 83(288), 2033–2060 (published online 18 November 2013). 10.1090/S0025-5718-2013-02787-1 · Checked 14 Oct 2026
- Tao, T. (2014). Every odd number greater than 1 is the sum of at most five primes. Mathematics of Computation 83(286), 997–1038 (published online 24 June 2013). 10.1090/S0025-5718-2013-02733-0 · Checked 14 Oct 2026
- Helfgott, H. A. (2014). The ternary Goldbach problem. In Proceedings of the International Congress of Mathematicians, Seoul 2014, Vol. II, 391–418. Kyung Moon Sa. Source · Checked 14 Oct 2026
- Bordignon, M. & Starichkova, V. (2024). An explicit version of Chen’s theorem assuming the Generalized Riemann Hypothesis. The Ramanujan Journal 64(4), 1213–1242. 10.1007/s11139-024-00866-x · Checked 14 Oct 2026
- Bordignon, M., Johnston, D. R. & Starichkova, V. (2025). An explicit version of Chen’s theorem and the linear sieve. International Journal of Number Theory 21(10), 2497–2572. 10.1142/S1793042125501192 · Checked 14 Oct 2026
- Li, J. & Liu, J. (2026). Theorem (1 + 1.9) on the Goldbach Conjecture. arXiv:2606.05224 (v1 1 June 2026, v2 8 August 2026). Preprint, not peer reviewed. Source · Checked 14 Oct 2026
- Helfgott, H. A. Publications (author's web page, last modified 19 December 2025). Source · Checked 14 Oct 2026
- Wikipedia. Landau’s problems (revision of 3 June 2026). Source · Checked 14 Oct 2026
Source
Open, as listed by Wikipedia · Landau’s problems. Checked 29 Sep 2026.
The status is the source’s. Papex does not decide whether a problem is solved.
Records on this problem
No public record addresses this problem yet.