'Repeatable theoretical model' is not an episemological criterion of a science, usually for empirical sciences what is needed are fasifiable predictions.
Falsifiable predictions are generated by theories. You don’t just throw out predictions haphazardly and test them, they always originate from a theory.