You're talking about the multiple checks of the same code runner question in the same quiz attempt correct?
I've assumed it's expected behaviour. Personally I prefer it that way, as often students are still tinkering with their answers when the test time runs out. If that happens while they're writing a line of code, they shouldn't get 0/1.0 or less than what did have working.
If it gave only the last result at submission, I would be needing to manually adjust a lot of questions.