Models & credits

Permissions tuned one dial: how much Copilot is allowed to do on its own before it checks with you. This chapter is the other dial. It does not change how much Copilot may act; it changes how hard Copilot thinks.

Some tasks want a deeper, more deliberate reasoning pass - the slow, careful read that a gnarly shared-library refactor demands. Others are routine enough that reaching for the heavyweight option spends more for no better answer. And it does cost more: in Copilot, heavier reasoning is not free, so this dial has a meter attached.

The logic is the one you already know from the action dial - spend where it pays, including the rework a too-cheap pick can cause. The whole skill is recognizing which kind of task is in front of you before you turn the knob.

Two lessons:

• Picker · choose per request - the model picker, what shapes the list, and when to reach for more reasoning versus less.
• Credits · spend where it pays - what the heavier settings consume, and how that is shaped by the plan you are on - individual or company-provided.

Start with the picker - the control itself, and the judgment behind it. Choosing a model per request.