Most people run their most powerful model on everything. Big tasks and tiny ones. That's what burns through your usage fast, and locks you out right when you need it most. The fix: match the model to the job. Same plan, same price, way more usage.
Model names change with versions. The families stay the same. Look for these in your dropdown.
Mine is built around my work. This prompt builds one around yours.
You are my model coach inside the Claude app. The big idea: Claude gives you a few different models, and if you run the smartest, slowest one on everything, you hit your usage limit fast and get locked out right when you need it. The fix is to match the model to the job... save your best model for work that matters, use the lighter, faster ones for simple stuff. I'm using the normal Claude app, not the coding tools, so we can't build files or automatic pop ups. Instead, I want you to learn what I use Claude for, give me a simple cheat sheet of which model to pick for each job, and write me a short instructions block I can paste into a Claude Project so you'll remind me going forward. Keep everything friendly and simple. Confirm with me before finalizing. How my plan works, so your advice fits. I'm on a Claude subscription, so I'm not paying per word, I have a usage limit that resets on a timer. I switch models using the dropdown at the top of the chat. Step one. Ask me these, in one short message. What do I use Claude for. Which of those is most important, the stuff I can't afford to get wrong. Which Claude plan am I on, and which models show up in my dropdown. Step two. Once I answer, match each of my tasks to a model and show me a clean cheat sheet. Step three. Write me a short Project instructions block I can copy, that tells Claude to remind me which model fits each task going forward, without nagging. Step four. Tell me, in three short steps, how to use it. Keep it simple. Plain English, no jargon. Short cheat sheet I can screenshot. Friendly reminders, never naggy.
This is roughly what Claude will write for you in step three.
At the start of each task, quickly tell me which model fits this job based on my cheat sheet. If I'm using a heavier model than the job needs, say so in one friendly line and remind me I can switch with the dropdown, then continue. Don't nag, mention it once.