reduce accidental MAX Mode.
easiest change: show when its enabled (currently hidden, have to click on the model)
small feature: add a notification or confirmation
small feature: show live token spending so we can learn how to minimize it.
complete solution: let us limit/throttle each request so we can make sure we don’t mistakenly have the most expensive agent do something extremely costly when its not important.
Right now we have Spend Limits for individual plans, including Ultra. You can set a monthly limit for on-demand usage in Dashboard > Spending. That said, this is about total monthly spend, not per-request control like you described.
Your suggestions about MAX Mode visibility, like a visual indicator showing it’s active, and per-request throttling make sense. Right now it’s really not obvious when MAX Mode is on. I’ll pass this feedback to the team.