I haven’t gone through all the code, but in this case the answer doesn’t depend on your particular code. The same computations would be performed under the hood either way, the only difference between model and transformed parameters is whether the variables are saved or not. So I think you’re right to have them in the model block if you don’t want them in the output.
If you want help looking for potential speedups to your model feel free to open another topic about specific parts of your code.