Get the most out of Claude Max and ChatGPT Pro

What you'll see in this lesson

1. How your limit is measured in Claude Max

   Max 5x and Max 20x give five and twenty times Pro's usage per session, with a five-hour window and a weekly cap.

2. ChatGPT Pro: three tiers and a shared pool

   Pro is sold in three tiers, at 100, 200 and 500 dollars a month; in Work and Codex all models draw from the same pool.

3. What eats up your limit

   It's not just what you type: history, attachments, tools, the model, effort and multi-step tasks all weigh in.

4. Choose the model for the task

   The model is the biggest lever on usage: at API prices, Opus costs twice as much as Sonnet and Sonnet about 20 times more than Haiku.

5. Effort: only raise it when the task calls for it

   More effort gives more thorough answers, but takes longer and uses more tokens: choose the level per task, not out of habit.

6. In ChatGPT, the model multiplies your messages

   On the same plan, choosing the light model gives you dozens of times more messages than the most powerful one, and fast modes cost extra.

7. Several sessions at once without draining the pool

   Working in parallel multiplies what you get done, but every session, subagent or teammate draws from the same limit.

8. Watch the meter: /usage and Settings > Usage

   Check how much you've used before launching something big, and find out what eats your allowance with the /usage breakdown.

9. Summary

   A high-tier plan goes twice as far if you choose the model and effort well and keep an eye on what uses it up.