Drive-test data scale of NR RLC throughput into NR PHY throughput

I’m trying to scale drive-test data of NR RLC throughput into NR PHY throughput.
Each vendor used different DT tools and provided me data in different layers.
For example, for LTE PHY is about ~24% larger than LTE RLC, already accounting all overheads and BLER.
Anyone have such reference for NR?

Hope this table will help:

Values are aggregated/binned per second so there may be weird cases when MAC Throughput > PHY Throughput by small values.

Should not consider physical layer throughput as it has direct dependency on BLER which is variable and depends on radio condition.
Better to look PDCP layer throughput which is muh closure to application layer.
In case of ENDC, we can check RLC throughput to check LTE and NR contribution.

It is a good suggestion, but for the analysis I was performing I needed them separated per RAT at the same layer, not combined/aggregated at PDCP or Application layers.