Your comparison it unfair.
OK, setting aside a "normal" portion of unbalance load current on the GEC, let's look at only connections along and in parallel with a grounded service conductor only between the service disconnect and the meter.
We're talking about connecting a conductor in parallel with the neutral that is not required.
What if I mount a service center with a metal enclosure beside a meter socket, and support them commonly on metal structural channel. Now, I could choose a non-conductive support, but I choose strut. The strut, of course, is conductive and creates a parallel path for current on the grounded service conductor. Is this a violation of 250.6? IMO, no.
Let's say I run the service conductors that go between the service center and the meter in a metallic raceway. This is almost always a design choice, only. The metallic raceway is a parallel path for current on the grounded service conductor. Is this a violation of 250.6? IMO, no.
Let's say the metal strut is mounted to a wall that is framed with metal studs and contains metal ductwork for the HVAC system and the metal ductwork is connected in places to structural building steel that in turn is connected to the earth (as well as the GES) which is connected to the PoCo transformer ground which in turn connects to the grounded conductor at the transformer secondary terminals. There's a nice parallel path here. . . 250.6? IMO, no.
I could continue, but, my point is the current is "objectionable" when it affects "sensitive electronics," not because its one of the many currents in the multitude of parallel paths for a grounded service conductor existing on the line side of the Main Bonding Jumper.