If Fable gets correct answer quicker, then you might pay less than doing back and forth with Opus, plus you lose more of your own time.

I see no reason for using less able models in my workflows. There is this saying, penny wise and pound foolish