Is my understanding correct:
Auto is a bucket seprate from API usage. After auto/composer is used, it will be billed at API usage and county against the API dollar amount of our plan, which has a minimum and some bonus variable usage.
Is my understanding correct:
Auto is a bucket seprate from API usage. After auto/composer is used, it will be billed at API usage and county against the API dollar amount of our plan, which has a minimum and some bonus variable usage.
yep! you got it right exactly
Yes, your understanding is correct.
Detailed Explanation:
Separate Buckets: As shown in your screenshot, “Auto” and “API” are tracked as separate buckets.
Auto: Tracks usage when you let Cursor choose the model (e.g., in Composer 1.5 (price), codex, etc ).
API: Tracks usage when you explicitly select a specific “named” model (e.g., manually selecting Opus 4.6).
The reason “Auto” is metered differently is that it optimizes for efficiency. Cursor dynamically selects the best model for each specific question.
For a simple query, it might pick a cheaper, faster model.
For a complex coding task, it might pick a more expensive reasoning model.
Therefore, the consumption is variable—you are billed for the actual model Cursor chose to use for that interaction.
API Pricing: When you bypass Auto and select a specific model yourself, that usage goes directly to the “API” bucket and is billed at the standard Models | Cursor Docs view prices.