Hi @uzytkownik1
The token usage depends on how much context you add to your request. This includes attachments, rules, MCPs,… but also how you use Chat, which models you use and if you use Auto model.
- More context: attached files, rules, MCPs,… means more tokens used in chat and with every chat follow up question
- Long chat: too long chat with a lot of context for unrelated tasks continues to consume tokens it does not need.
- Heavier models: the more capable but also consume more resources.
As there is no simple formula that fits everyones usage since tokens cost differently per token type, provider and model, please use following info.
Here is more information and best practices to reduce token usage and optimize cost.