This idea that you need to model GR to do GNSS is very over stated. You need to correct for a GR effect, but it is a small part of a residual in a least squares

This idea that you need to model GR to do GNSS is very over stated. You need to correct for a GR effect, but it is a small part of a residual in a least squares problem that gets regressed away. So it is important for understanding the error budget but it really is a small effect, compared to say ionospheric delay due to atmospheric charge or clock drift.