Here's a way - not the most efficient, but it works:
sed 's/\(.*[^0-9]\)\([0-9]*\)/\1 \2/' input_file | sort -gk2 | sed 's/ //'
It assumes a quantity of alphabetical characters, followed by digits.
colucix' idea above is very nice (thanks for that colucix!) and is something I didn't know about with `sort` - but as he mentioned, it is strict on the number of characters.