Archiving (grouping files in a single entity) and compressing are usually done in separate steps under Unix.
That gives more flexibility when you want to manipulate single files (that wouldn't benefit from archiving) — think pipelines such as
gzip -cd mylog.gz | grep '^Oct.*error'. Achieving the same with
7z would be awkward.¹
arj, and...) are archivers that happen to compress. They wouldn't replace traditional
gzip as they do not have the same purpose.
gzip is fast, light on resources, very well-known and installed on most Unix systems. If you're looking for a higher-performance compressor, go for
xz. It offers a command line interface very similar to
gzip but uses more or less the same LZMA algorithms as
7z so you should get similar compression ratios.
1) The above pipeline is so common that many wrappers exist around
gzip tools; a seasoned user would probably opt for
zgrep '^Oct.*error' mylog.gz to save a few keystrokes.