In support of the previous answer, grouping zones by similarities, occupancies and/or orientation is a customary practice and is a suggested approach to keep model complexity (and subsequent runtimes) as low as possible.
A useful means of thinking about zoning may be to identify the major energy consuming components of the design to make sure they are adequately reported and tested against the Standard and zone accordingly.
In the case of High-rise res, lighting is not a factor within the residential occupancy (only non-res spaces like lobbies, corridors, etc. are tested) leaving Envelope, DHW and HVAC.
Skipping Env and DHW for now.
Knowing that CBECC-Com is very granular in the testing of fan power and load right-sizing, grouping zones by System/Load (as noted in a previous answer) is a good means of maximizing what credit the simulations will offer you (especially knowing there will be little lighting credit to assist in compliance in the case of res projects).
Example: Corner Suites with multiple wall orientations (and glazing) would be modeled separately from those Apts with only one wall orientation only to be further broken down and zoned by the size of the HVAC Unit.