Since composer is like over double the price of auto. (assuming api fees)
it sure seems to me that Auto uses up the Auto + Cursor allocation at the same rate.
_And_ to me it seems like auto consumes less tokens per request.
Anyone else notice this?
It’s hard to know for sure if Auto uses fewer tokens, because we don’t know which models it’s using under the hood
i could argue that its at least fairly similar usage. its just so difficult to tell if you are using your auto/composer percentage faster.. there should be a over 50% difference.
We added extra usage for Auto / Composer 1.5 pool, but that is not related to how many tokens each model consumes.
As most models are reasoning models, the difference is in their trained ‘behavior’ like being helpful, proactive etc.
So within the Auto / Composer 1.5 pool do you get roughly twice as much usage on Auto over Composer as they have a different price or is it based on token usage count only?