“This one feels better” isn't a reason. Put your criteria in with weights, score both responses out of five, and get a weighted result and the exact criteria that decided it — useful for picking a model, a prompt version, or settling an argument.
Nothing saved yetYour inputs are kept in this browser only — nothing is sent anywhere.
Rubric weight is relative; score each response 0–5