Hey Product Hunt,

85% of my weekly Claude limit was gone by Wednesday. When I looked at why, most of it was the top model doing easy work: rewrites, summaries, quick questions a lighter model handles just as well.

The hard part is knowing which model is right on the plan you actually pay for. Every app has different models, thinking settings and limits, and they change every month.

So I built Tierpik. Type what you're about to ask, pick your app and plan, and it tells you the exact model and setting to use, plus when it's worth stepping up.

- Works for ChatGPT, Claude, Gemini, Grok, Perplexity and Copilot, Free to top tier

- Developer mode picks the cheapest API model for the job

- There's a connector so you can ask inside Claude

- Free, no sign-up, runs in your browser

It's part of a bigger experiment: 10 genuinely useful free apps, built cheap and kept running.

What do you use the top model for that a lighter one could handle?