Three dials, one bill
GitHub gave Copilot’s model picker three settings, and the bill ignores which one you pick
Efficiency, balance and intelligence draw on the same set of models, and charging follows whichever model answers.
GitHub added three tiers to Copilot’s automatic model selection on 14 September, named efficiency, balance and intelligence.
The tier is a preference, not a route. GitHub’s entry says “all three tiers use the same set of available models” and that auto “individually evaluates each prompt”, offering its own example: a docstring may still be written by a small model under the intelligence setting.
So the setting cannot be used as a cost control either way. GitHub is explicit: “Usage is charged based on the model auto selects, regardless of tier.” Paid subscribers keep a tenth off usage billed through auto.
What the entry never says is which tier answers for the people who never open the setting, and GitHub calls the feature “the first step”.
What it changes Do not use the tier as a budget control; it changes a preference and not the price. If you track Copilot spend per seat, keep tracking the model names in the usage export rather than the tier.
Primary source read