Get estimated sampling time for a model as a variable?

I have a model I would like to fit to various datasets, some (but not all) of which are too big to sample in a reasonable time without using ADVI (as opposed to NUTS). Where possible I would like to use NUTS in order to get a more accurate estimate of the posterior. Is there a way to get the estimated time to generate a NUTS sample for a given model so I can use ADVI instead if NUTS would take too long?