Or you use the fmpy command compile_platform_binary (also on the target platform) as can be seen in this example: https://github.com/CATIA-Systems/FMPy/blob/main/tests/test_c_code.py
Regarding Intel vs. M-chip: I am not familiar with this. For FMI 3.0 this should generate different binaries with different platform tuples, see https://fmi-standard.org/docs/3.0.2/#platform-tuple-examples. But I never tried this on a Mac.