You seem to have a concept that a scientific model is either "true" or "wrong", but that's not case. All models are wrong, but some are more useful than others. It's better to judge models by their ability to describe reality, and that is not a binary property, but a continuous one.