How many treatment arms should I create when the same treatment arm is part of both blinded and open-label periods?

Create only one treatment arm. The unblinding setting on the randomization design determines whether the treatment arm title is visible to blinded users.