After I tried to roll back to the old pricing plan, during the process of using the Sonnet 4 model, I found that whether I interrupted the continuous output or resent the request, it would rudely deduct one request count directly.
When I switched to the auto - mode, for some complex problems that required associative thinking and solution - seeking assistance, it would only list plans for me like the old model (without automatic execution), and write code for me (without automatic application). This forced me to give up the auto - mode.
Based on the above - mentioned situations, is there a way to optimize the technical mode to allow me to maximize the utilization of each of my requests? I need help.