Pool vs BasicVariantGenerator for multi-classifier tuning

I want to tune multiple classifiers in the same script.
For this use case, is it recommended to make a list of dictionaries with classifier and configs and use Pool to tune them in parallel?

or

Is BasicVariantGenerator already has an implementation for this? I see that we can give multiple configs but not sure how to give the classifier to Tune.run or TuneSearchCV ?

or

Do you recommend a completely different way that is more effective than the above two?

Please suggest!
Thanks!