Nonmem to Mrgsolve validation

[Ravua1992]

I converted a 2‑compartment model from NONMEM to mrgsolve. SC (subcutaneous) route results match NONMEM PRED, but IV post‑infusion concentrations at the end of infusion do not match (they are higher compared to Nonmem PRED).
I verified input parameters and model structure; they appear consistent between NONMEM and mrgsolve.
I’m looking for suggestions where to check (model/code/solver/settings, infusion related settings.) and for help identifying what might cause these differences.

I am following the steps listed in this post:

https://mrgsolve.org/blog/posts/2022-05-validate-translation/

Nonmem PRED vs Mrgsolve PRED (SC= ROUTEN 3 and IV= ROUTE 2 ):

Summary between nonmem and mrgsolve pred:

summary(summary_df$IPRED - summary_df$PRED)
     Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
-33.66590  -0.71977  -0.20104   0.79590  -0.02835 212.45660 

Thank you