Question

Is a model with a high R-Squared value always better than one with a low R-Squared value?

Answer 1

If all assumptions of the models are verified, yes

Explanation:

The R-squared value is the amount of variance explained by your model. It is a measure of how well your model fits your data. As a matter of fact, the higher it is, the better is your model.
However, it only applies when te assumptions of the models are fulfilled (e.g. for a linear regression : homogeneity and normality of the data, independence of the variables, fixed X).

For exemple, the figure below represents four models with exactly the same R-squared. As you see, only the first one seems to really fit the data.