A little shorter maybe and doesn't require any modules...

sub longest_words {
    my @W = $_[0] =~ /\b(\w+)\b/g;
    my $L = 0;
    $L = length() < $L ? $L : length for @W;
    grep length == $L, @W;
}
[download]

I wouldn't actually do it this way though. I'd pay more attention to doing it in a single pass through the list of words.

Update: Switched back to the regex for extracting the words.

-sauoq
"My two cents aren't worth a dime.";