Model Settings

Token count: 0
Higher effort spends more output tokens on reasoning before replying.