It's often recommended to put the cells in parallel to initially charge/balance them. I did this with some headway cells, then assembled them into a series/parallel pack, left them for a month or so and came back, there was a 0.05v difference between a few of the groups, however after putting a light load on them (5amps) they instantly became the same voltage under load, I left the load there for a couple hours and every group stayed the same. I wouldn't worry about the difference, but it's always a good idea to get them to the same level before use, either top/bottom balancing depending on the method you believe in.
The other thing you could check is swap the BMS from the 3.2v cell to a different one perhaps it's drawing more current for some reason?