Yes, you can choose freely from any model
that has been transformed by orthonormal rotations. Remember that the systematic features captured by the models are the same and that the fit is unaffected.
commands to get an overview of the two matrices.
By now you should have experienced
that there can be (great) differences between cores from rotated models. And as your interpretation always will start by identifying the significant factor combinations, your highest desire will be to get a core that contains a few but very large elements since this will give a clear and direct approach to understanding the most important features of the model, hence, the systematics of data.
Thus, if you have more models to choose from you should select the one with a straight forward interpretable core.