question on z-buffer depth test algorithm

hello, i was wondering :

inside the rendering pipline, after the projection transformation
all the vertices have the same z value (aren’t they ?)
so how can the depth test compare depths of different fragments ?

does it somehow holds the vertices depth from before the projection ? maybe some sort of interpulation betwen them ?

the reason i ask is that i was given this exercise in a course im taking on computer graphics :

it starts with calculating a projective transformation on a given 3 vertices and then
i am asked to describe how the z-buffer whould look like
after rendering these vertices as a triangle.

all help would be much appreciated !

clearification:

this exercise was given for self practice before the test
it is not obligatory and will not be submitted

so you are not helping me cheat in any way if that bothers anyone

all the vertices have the same z value (aren’t they ?)

Not at all. And they even have different w values.