If you don't care about order, replacing the deleted element by the last in the array works as well. And I was expecting it to be faster than splice (since I was expecting the latter to move every element after the one that is removed) but the performances are actually very similar:

use warnings;
use strict;
use Benchmark qw/cmpthese/;

my $size = 1e3;
my @numbers = ( 0..$size );
my $index = 3;
my @expect = ( 0..2,4..$size );
use constant TEST => 0;

cmpthese(-2, {
  splice => sub {
    my @output = @numbers;
    splice @output, $index, 1;
    join("\0", sort @output) eq join("\0", sort @expect) or die if TES
+T;
  },
  grep => sub { # https://www.perlmonks.org/?node_id=11123877
    my @output = @numbers[ grep $_ != $index, 0 .. $#numbers ];
    join("\0", sort @output) eq join("\0", sort @expect) or die if TES
+T;
  },
  swap => sub {
    my @output = @numbers;
    $output[$index] = pop @output;
    join("\0", sort @output) eq join("\0", sort @expect) or die if TES
+T;
  },
});
[download]

C:\Projets\perl>perl pm_select_11123879.pm
          Rate   grep   swap splice
grep   14287/s     --   -73%   -73%
swap   52519/s   268%     --    -0%
splice 52549/s   268%     0%     --
[download]

NB: I checked, it works fine when $index is the last element in the array as well.