What kind of classification problem are you working on? Might help narrow down which metrics make the most sense for you.
But generally First up, you’ve got accuracy. It’s like, how often did your model get it right? Simple, but sometimes it can be misleading if your data’s skewed.

Then there’s precision and recall. Precision is like, when your model says “yep, that’s the thing,” how often is it actually right? Recall is more about how many of the actual things your model caught.

F1 score’s pretty handy too. It’s basically a way to balance precision and recall. Useful when you don’t want to favor one over the other.