Cost for higher token usage

[cralicriux]

Let me start off with, this project is amazing! I'm still very new to Claude code and PAI. I've been using opencode for a few weeks for AI in general, dipping my toes into an AI assistant, but with Opencode I haven't had to see how much tokens I'm using really.

But let's get to the point. It seems that, and makes sense too, that PAI is using much more tokens. Has any thought been put into having different models/modes to have PAI (automatically determined based on the tasks) run in to better keep token usage down? I've already burned through my Claude pro tokens each day within about an hour :P

Right now I'm thinking to manually just switch the model as I'm using PAI, but thought it would be cool to have PAI set up so that certain low level functions use lower level models to save costs/time.