If the camera characteristics are different then 2 different camera models are produced. If they are the same then just one will be produced, so it shouldn’t be a problem.
I’ve combined images from 2 different M2Ps, which one camera model was used, and also phone and DSLR, where 2 models were produced.
I saw in the last versions that is also possible that the model has problems if you put in pictures from the same camera but with different zoomlevels. For example from the Mavic 2 Zoom. Never saw that this worked Somewhere.
For sure: if the software doesn’t know you’ve used different zoom levels, it can’t derive a camera model. To work around this (if you must use a zoom camera, which have their own challenges), re-label the model name in the images for each of the zoom levels.