> python2's implementation of randrange() that uses random() 
> under the hood was noticeably faster than python3's 
> randrange() that uses getrandbits() under the hood.

Yes, that was a conscious decision. See https://bugs.python.org/issue9025 . We traded performance in order to gain correctness.  As noted in the docs:

"Changed in version 3.2: randrange() is more sophisticated about producing equally distributed values. Formerly it used a style like int(random()*n) which could produce slightly uneven distributions."