Hi Team,

Hi Team, I have many many difficulties to understand agent's token consumption. Some things are not understandable (like using our knowledge in our Space is 3 credit cost -> so we put our knowledge in pods that cost 1 credit...) I struggle with the notion of credits: I am unable to pick the right model for the right task. I don't have enough information; I use some models for their effectiveness on certain tasks (as reading long documents, or data, etc) but I can't estimate their cost. And I can spend 4000 credits on one big task with subagents. I think we partners have a special need to pilot our client's bills - We have angry clients and it's not good... We promise to ajust their consumption and it's very hard to make a fine-tuning with so few cost information. Thanks