Ogden and Mooney-Rivlin (along with all of the other hyperelastic models) are merely abstractions used to describe the material. Given that, the "best" is the one which provides the response which is most similar to your particular material and its particular operating range.
If there was one superior model for hyperelasticity, the others would no longer be relevant.
The best recommendation is to obtain material data, fit it for all available models which your software can model, and choose the most appropriate fit. You may want to refer to thread727-53543 which describes some of this in much more detail.
Brad