GPS in general is a time based - better said transit time based - system.
You need precise information about time and position of located satellites and a really precise clock inside the GPS receiver. Normal quartz clock technology is not precise enough.
The GPS signal includes a clock signal that allows the GPS receiver to synchronize to the GPS clock and calculate position if at least 4 satellite signals are received. This makes the receiver much cheaper and portable.
Otherwise you might need an atomic clock
Now to the 5D4:
The camera has an internal quartz clock that works independent from the GPS system.
It seems that this one is not synchronized with the GPS time generated inside the GPS receiver.
Maybe there is a synchronization function included but not activated, or maybe it's just not there.
I don't know, which one of these two time signals is written into which part of the exif data but you'll surely find that information by internet search.
So you have two independent time signals. And depending on the quality of the quartz those two time values can differ within a short time period. I would expect a few seconds within a few weeks.
If it's more than a minute within a month IMO you should refer to Canon if that is normal.
I hope that helps a bit.
By the way:
What is the third clock you mentioned inside the body?
Edit: I did a quick search in the 5D4 manual, didn't find any "third clock" but an "Auto-update" function for synchronizing the internal clock with the GPS clock. Refer to the GPS chapter and/or the GPS menu of the camera.