I'm on the Pro plan for Claude 4 and not using Claude Code, opting for MCPs and Commander for my coding tasks instead. With my programming background, I've been testing models on my project, but I keep hitting the message limits way too quickly—just five messages in Sonnet before I had to switch to Opus, and then I hit the limit after one message! I want to make my prompts more efficient because it's frustrating getting locked out for hours due to hitting these limits. Right now, I'm wary about using Opus 4 since it consumes tokens so rapidly. I'm considering sticking with Sonnet, and I might even stop using Commander to update my files more directly, even though it's slower, to help with token usage. Do you have any strategies or tips to help me maximize my usage without upgrading to the Max plan?
1 Answer
Using MCPs can really chew through your tokens without mercy. Try breaking your code into smaller chunks when you submit them. It's tough with the limits, but smaller requests can help you avoid hitting caps too quickly. Plus, once you hit a limit, you lose access to all models, which is a bummer. So pacing is key!
I feel you! It's frustrating when you reach your limit and can’t even switch up your model. It used to be easier to juggle them, now it feels like a balancing act. Gotta be strategic!