betaFeedback

WorksMathematics

A narrower admissible 234 360 000-tuple and H_5 <= 4 891 341 734

Abstract

Let H_m = liminf (p_{n+m} - p_n). Kapila proved DHL[234360000,6] and deduced H_5 <= 5040082836, the diameter of the tuple formed by the 234360000 smallest primes larger than 234360000. We construct an admissible 234360000-tuple of diameter 4891341734, which gives H_5 <= 4891341734 with Kapila's theorem. The tuple consists of +-1 and of +-q for primes q in a symmetric window, with 1746831 primes removed; its admissibility is checked by two independent programs.

Result

For m≥1m \ge1 let Hm=lim inf⁡n→∞(pn+m−pn)H_m = \liminf_{n\to\infty}(p_{n+m}-p_n), where pnp_n is the nn-th prime. A set H\mathcal{H} of kk integers is admissible if for every prime pp it misses at least one residue class modulo pp; it suffices to check the primes p≤kp \le k. The statement DHL[k,m+1]\mathrm{DHL}[k,m+1] asserts that every admissible kk-tuple has infinitely many translates containing at least m+1m+1 primes, and it implies Hm≤diam⁡HH_m \le\operatorname{diam}\mathcal{H} for every admissible kk-tuple H\mathcal{H}. Following Maynard [2], the Polymath project [3] obtained H5≤80 550 202 480H_5 \le80\,550\,202\,480, Stadlmann [6] improved this to 77 510 685 23477\,510\,685\,234, and Roeling [5] to 14 541 349 28814\,541\,349\,288. Kapila [1, Theorem 1.1] proved DHL[234 360 000,6]\mathrm{DHL}[234\,360\,000,6] and obtained H5≤5 040 082 836H_5 \le5\,040\,082\,836. This number is the diameter of the tuple of the k=234 360 000k = 234\,360\,000 smallest primes larger than kk [1 Table 3], which is admissible because none of its elements is divisible by a prime p≤kp \le k; we recomputed this diameter exactly (ancillary file previous_record_width.py).

Theorem 1. There is an admissible 234 360 000234\,360\,000-tuple of diameter 4 891 341 7344\,891\,341\,734.

Corollary 1. Assume DHL[234 360 000,6]\mathrm{DHL}[234\,360\,000,6], which is proved in [1]. Then H5≤4 891 341 734H_5 \le4\,891\,341\,734.

The only new ingredient is the tuple: the bound relies on the statement DHL[234 360 000,6]\mathrm{DHL}[234\,360\,000,6] of [1], which we use as stated.

The tuple

Let P0=50 100P_0 = 50\,100, a=2 445 670 867a = 2\,445\,670\,867 and let RR be the set of the 1 746 8311\,746\,831 primes listed as successive differences in two ancillary files, to be read one after the other: removed_gaps_1.csv, then removed_gaps_2.csv; they all lie in (P0,37 847 053](P_0, 37\,847\,053]. Put

A={1}∪{q prime:P0<q≤a}∖R,H={±m:m∈A}.A = \{1\} \cup\{q\ \text{prime}: P_0 < q \le a\} \setminus R,\qquad\mathcal{H} = \{\pm m:m \in A\}.

Then ∣H∣=234 360 000|\mathcal{H}| = 234\,360\,000, and aa is prime and lies in AA, so diam⁡H=2a=4 891 341 734\operatorname{diam}\mathcal{H} = 2a = 4\,891\,341\,734.

Lemma 1. H\mathcal{H} is admissible if, for every prime p≤234 360 000p \le234\,360\,000 with p>P0p > P_0 and p∉Rp \notin R, there is an integer cc with 1≤c≤20 0001 \le c \le20\,000, 2c<p2c < p, such that no m∈Am \in A satisfies m≡±c(modp)m \equiv\pm c \pmod p.

Proof. If p≤P0p \le P_0, no element of H\mathcal{H} is divisible by pp, since the elements are ±1\pm1 and ±q\pm q with qq prime and q>P0q > P_0. If p>P0p > P_0 and p∈Rp \in R, the only element of {1}∪{primes}\{1\} \cup\{\text{primes}\} divisible by pp is $p itself, which is not in AA; so the class $0$ is missed. Otherwise, by hypothesis no element ±m\pm m of H\mathcal{H} lies in the class cc mod pp, because m≡cm \equiv c or −m≡c-m \equiv c would give m≡±cm \equiv\pm c. For p>234 360 000=∣H∣p > 234\,360\,000 = |\mathcal{H}| some class is missed trivially. □

How RR was chosen. We started from {1}∪{q prime:P0<q≤2.5⋅109}\{1\} \cup\{q\text{ prime} : P_0 < q \le2.5 \cdot10^9\} and treated the primes p≤kp \le k with p>P0p > P_0 in decreasing order: we searched the smallest c≤20 000c \le20\,000 with both classes ±c\pm c mod pp empty, and if there was none we removed pp. Removing elements keeps the classes found earlier empty, so the final set satisfies the hypothesis of Lemma 1; H\mathcal{H} consists of the elements of absolute value at most aa. This is a variant of the sieve of Hensley and Richards, which sieves a symmetric interval [−x,x][-x,x] by the classes 00 mod pp for increasing primes pp (see [4]); the difference is that for most primes p≤kp \le k we do not sieve the class 00 but exhibit an empty pair of classes ±c\pm c, so that these primes ±p\pm p stay in the tuple. The removed primes are those for which the expected number of elements per class, about ∣H∣/p|\mathcal{H}|/p, is too large for an empty class to exist or to be found within the search budget; they are all smaller than 3.8⋅1073.8 \cdot10^7.

Verification

  • verify_tuple.py (a C program embedded in a Python script, compiled with gcc) builds AA with a sieve of Eratosthenes on bytes, removes RR, checks ∣H∣=234 360 000|\mathcal{H}| = 234\,360\,000 and a∈Aa \in A, and then, for every prime p≤234 360 000p \le234\,360\,000, scans the class 00 directly (all multiples of pp up to aa) and, if it is occupied, searches a pair ±c\pm c with c≤20 000c \le20\,000 that is empty, again by direct scanning. It reports 12 868 67312\,868\,673 primes, the class 00 missed for 1 751 9741\,751\,974 of them and a pair ±c\pm c for the other 11 116 69911\,116\,699, and the diameter 4 891 341 7344\,891\,341\,734. It needs about 3 GB of memory and three minutes.

  • build_tuple.py reproduces the construction and writes the list of the classes cc found for each prime (about 178 MB, not included). verify_classes_numpy.py, written independently in Python with numpy, rebuilds AA with its own sieve and checks this list: every prime P0<p≤kP_0 < p \le k has exactly one entry, every listed pair ±c\pm c is empty, and the cardinality and the diameter are as stated.

  • For k=111 300k = 111\,300 the same construction gives a tuple of diameter $1,449,286$; both programs accept it, and a direct computation of all residues modulo every prime p≤kp \le k confirms its admissibility. verify_tuple.py rejects the tuple in which one removed prime is put back and another element removed (no class is then empty modulo 1021, by exhaustive search), and verify_classes_numpy.py rejects a certificate in which one class is altered, both for k=111 300k = 111\,300 and for $k = 234,360,000 in the second case.

Remark 1. The diameter is about k(log⁡k+1.60)k(\log k + 1.60). For the values of kk used in [1] for H2,H3,H4H_2,H_3,H_4 (2 355, 111 300, 5 129 500) the same construction gives the diameters 25 562, 1 449 286 and 90 027 092, larger than those of [1]; so we do not improve these bounds.

Acknowledgements and use of AI

The mathematics, the computations and the writing were produced with Claude (Anthropic) under the author’s direction. The bound depends on the statement DHL[234 360 000,6] of [1]. The result has not been examined by a professional mathematician.

Independent researcher, Meaux, France

URL: https://orcid.org/0009-0009-1526-1178

References

  1. [1]A. Kapila, Improved bounds on Hₘ for 2 ≤ m ≤ 5, Hexagon preprint, hexagon:2610.00419v1 (2026).
  2. [2]J. Maynard, Small gaps between primes, Ann. of Math. 181 (2015), 383–413.arxiv.org/abs/1311.4600
  3. [3]D. H. J. Polymath, Variants of the Selberg sieve, and bounded intervals containing many primes, Res. Math. Sci. 1 (2014), Art. 12; arXiv:1407.4897.arxiv.org/abs/1407.4897
  4. [4]A. V. Sutherland, New bounds on gaps between primes, lecture slides, https://math.mit.edu/~drew/PrimeGaps.pdf.
  5. [5]M. P. Roeling, NeuralCert: certified computational discovery of extremal mathematical constructions, arXiv:2609.30296 (2026).arxiv.org/abs/2609.30296
  6. [6]J. Stadlmann, On primes in arithmetic progressions and bounded gaps between many primes, Adv. Math. 468 (2025), Art. 110190; arXiv:2309.00425.arxiv.org/abs/2309.00425

Paper details

vlnc:W_s5qevf3r
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Anthropic Claude Opus

View statement
Looking for related papers…
Contents