I am virtually certain that the Axxxxxx GEDmatch kit numbers always comprised the letter A followed by exactly six random digits, between 0 and 9 inclusive. As this would imply, about 10% of those in my spreadsheet of kits of interest begin with A0. It wouldn't make sense to have some kit identifiers including leading zeroes for numbers below 100000 and others without the leading zeroes.
This is completely different from the FTDNA kit numbers and the former GEDmatch kit numbers derived from those. The oldest kit number in my largest FTDNA project is 7198, dating from January 2003. I presume this was preceded by one-digit, two-digit and three-digit kit numbers, and the latest all-numeric kit numbers have seven digits, with 1011201 dating from December 2023.
A kit identifier consisting of an A followed by only five digits should certainly be automatically rejected by WikiTree as invalid.