can i submit multiple implementations to see with which language or algorythm performes best?
If you are doing this for Long Challenge, your best solution is considered. Long challenge doesn’t considers time taken to solve a problem nor has any penalties for wrong submissions.
If you are trying it in Cook-offs or Lunch Time challenges, you might get a TLE with some languages or some algorithms. There will be penalty for wrong submissions in these contests. So, this effect in ratings for that respective challenge. But in long-challenge it doesn’t.
they all get rated and the points count for the best?
or just the last submission?
your best points will be counted