“Garbage in, garbage out” applies.
Always has, always will.
The bar for calling models any sort of science is clear and high.
The gold standard for judging the validity of climate models to predict weather is simple:
If the model can accept instrument obtained input from weather, say 10 years ago, and compare the model provided weather from 20 years ago for the same period as the observed weather 10 years ago, the model is valid.
No known model to date has ever passed this simple test.
Any claim to the contrary involves fraudulent "massaging of the facts"