The bulk insert will undoubtedly be much faster than using transforms, but an important question is what kind of transforms?  If you are modifying data in transforms, then these operations will have to be done in a separate UPDATE after the bulk insert.  If you need transforms to clean or correct data (illegal dates, formatting issues, etc.), the bulk copy will probably fail.  If the DTS package is running on a client PC, the transformation workload is on the client instead of the server, although you would need some heavy-duty transformation logic for this to outweigh the speed advantage of a minimally-logged bulk insert.

There are exceptions, but I would say use the bulk insert if it will get the job done.  Even if an UPDATE or INSERT/SELECT is needed to modify the data, this will probably run faster in T-SQL than in ActiveX transforms.  I can't say how much faster, there are too many variables.