All usage of ChatGPT is governed by a single credit usage meter. However, some daily chats consume much fewer tokens than long-horizon work tasks. I often run parallel chat threads and work tasks, while once the limit is hit, all usage are immediately terminated. Sometimes, the work can wait until the limit is reset to continue, but I often would want to continue some conversations that use few tokens.
Could there be a customizable usage management function, that reserves say 10% tokens by default for selected chats, while the bulky tasks can only consume up to the remaining 90% unless I authorize it to use the reserved usage?