Is Claude Max or ChatGPT Pro worth it?
Say how much of your weekly limit you use. The calculator applies the home page's estimates to it: what your week holds, what it would cost at API list prices, and how much of each plan's weekly limit it would take.
Your estimate
Your week: about 52M tokens.
At API list prices that is about $200 a month.
ChatGPT Plus (Codex) holds it, at 100% of its weekly limit.
| Plan | Share of its weekly limit | Your usage at API prices | Basis |
|---|---|---|---|
| ChatGPT Plus (Codex) | 100% | 10× its price | Measured |
| ChatGPT Pro $100 | 20% | 2× its price | Projected |
| ChatGPT Pro $200 | 10% | 1× its price | Projected |
Measured rows come from a run on that plan. Projected rows are scaled from a measured plan by the home page's assumed multiple and were not measured on that plan. Both labels describe the source of the capacity estimate.
How the calculator works
It starts from the home page cards for your plan, model and kind of work. Your week is that percentage of the card's weekly estimate, and its value at API list prices is the same percentage of the card's monthly API value. Treat the week as work on the selected model and kind of work.
The other plans' shares use the same multiples as the cards. How the numbers work explains each multiple and what was measured. It compares plans from the same provider; it does not convert one provider's weekly percentage into the other's.
The value at API list prices is what your weekly usage, every week, would cost a month at API list prices. The quota estimate comes from measured weekly-meter movement; API prices value the recorded token mix. API value is a list-price comparison; your actual API bill or saving can differ.
"More than" or a > prefix means the figure is a lower bound. Weekly shares use the same base run on both sides, so that bound cancels. Figures are shown to two significant figures.
What it leaves out
Your own mix of input, cached input and output can differ from the run's, so your week can hold more or less than the estimate. Five-hour, model and image limits can stop you before the weekly limit does. ChatGPT Pro $500, credits and extra usage are not in the cards.
To find your percentage on Codex, see how to check your Codex usage.
For each plan's rules, see Codex pricing and Claude Pro vs Max. To measure your own usage, install the token-counter plugin.
This site is not affiliated with, endorsed by or sponsored by OpenAI or Anthropic.