A 485-point spherical code in five dimensions with maximal cosine below 0.855765650858
There is a 485-point spherical code in R^5 (S^4) with maximal cosine 0.85576565085796... < 0.855765650858 (minimal angle 31.1556°). This improves the best known value 0.856463812595 for N = 485 in Cohn's spherical-codes table (spherical-codes.org; A. Jay, 2026), and was verified by an exact integer check of all 117,370 pairs. The table is Henry Cohn, Table of spherical codes, https://spherical-codes.org/ (archived data set: https://hdl.handle.net/1721.1/153543); the comparison uses the row listed on 18 November 2026. The previous 485-point code is due to Antoine Jay (2026), made with the program of A. Jay, M. Poberznik, M. Brut, A. Hemeryck, L.-H. Hoang, M. Gunde, N. Salles, G. Cruchon and M. Roux, "Equidistribution of points on the hypersphere including symmetries", preprint, 2026, https://doi.org/10.2139/ssrn.6427919 . The construction starts from our 484-point code (https://doi.org/10.5281/zenodo.24343765), itself built from our 483-point code (https://doi.org/10.5281/zenodo.24342564): one point is added at its deepest hole, then all 485 points are relaxed jointly on the sphere and polished with L-BFGS (scripts build_insert_relax.py and build_polish_lbfgs.py; they need numpy and scipy and are given for the record only). The final coordinates are integers; each row v of directions.csv stands for the unit vector v/|v|. Verification: run python3 own_check_exact.py directions.csv. It uses only the Python standard library and compares the exact squared cosines (v·w)^2/(|v|^2|w|^2) of all pairs with positive product as fractions; the maximum is attained at rows 115 and 144. The certificate does not depend on the numerical optimization. This is a construction at one cardinality. It says nothing new about the kissing number in dimension five, and the optimal angle for 485 points remains open. AI use: this note comes from an AI-assisted research lab directed by the author. The construction, the exact verification and the text were produced with Claude (Anthropic) under the author's direction, and the check was repeated by a separate session of the same AI system, so this is not independent human review. Not peer reviewed. Contact: [email protected].
Authors
- Учрал Ням-Осор
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-18
- DOI
- https://doi.org/10.5281/zenodo.24344670
- Primary Topic
- Mathematical Approximation and Integration
- Type
- preprint