Charmm22 or Charmm36 force field for homology models?

GROMACS version: 2023.3
GROMACS modification: Yes/No
What force field is better for simulation of homology modeled proteins? Two popular programs for homology modeling: MODELLER and Swiss Model Server use Charmm22 force field for parametrization for structures modeling.
So is it preferably to use Charmm22 force field or newer Charmm36 in Gromacs for molecular dynamics of such proteins?

CHARMM22 was considered state of the art about 20 years ago :)

Use CHARMM36. It has basically made old versions of the force field obsolete.

Thanks a lot