if you want different prune behaviour per backup job, you can configure the pruning client-side (either in the backup job, or on the storage, or on the node via vzdump.conf). but of course this only works if the server side pruning (if one exists) is more lax, else the server-side pruning will prune things that the client wanted to keep.
GC is strictly server-side and per datastore, it's not possible to do anything "per group(s)" there..