Which best captures the key idea of "Matching model to task"?

AGive constructive, specific feedback to students BSame questions, scorecards, evidence-based decisions — fair and legally safer CPick models by strength: reasoning, code, search, vision DRAG reduces hallucination and lets models answer about private, current data