4 ms·

Each GHA runner gets its own VM. So every minute those are running you're billed. The runners do work independently which can save wall time.