How do I disable the setting to automatically treat all subsequent agent prompts as max-mode prompts after I select max mode on a previous prompt? I enabled max mode on GPT 5.6 Sol, and every agent I opened was charged as a max-mode query, according to my usage dashboard. How do I disable that without resorting to disabling max mode on a model that offers it, starting a new agent chat/instance, and checking if it was billed as a normal query according to the usage dashboard?
If I select max-mode but the model consumes fewer tokens than the normal non-max token window, is that billed the same amount as a non-max-mode prompt/call? Cursor’s documentation AI assistant only states that there is an extra 20% fee for max-mode requests. Is that 20% fee applied even if the tokens did not exceed the context window? What is the benefit of max mode aside from a larger context window?