Hi,
Regarding the first question. No it is not the same model since you add a hyperparameter. Now your betas will be correlated. Besides in the second model you only have one “sigma” instead of several.
If the first model is too “wordy” why don’t you use a shape parameter for mus and sigmas?
Regarding the second question, I would not know exactly the syntax but you will have extended possibilities by using arviz instead of pm.traceplot