Pro 200 cut from 20× to 10×: how can customers verify what these multipliers actually mean?

Codex
1

The reduction of Pro 200 from 20× to 10× made me question what these multipliers actually represent.

I asked OpenAI Support what absolute quantity is behind the usage percentage shown in Settings → Usage.

Support confirmed that:

  • The quantity behind 100% is not disclosed as a customer-visible absolute unit.
  • It is not defined as a fixed number of messages, tokens, tasks or published compute credits.
  • OpenAI does not document a fixed absolute equivalent for 5×, 10× or 25×.
  • The documentation does not explicitly guarantee that 10× always means exactly ten times the same internal Plus baseline.

This is particularly relevant because Pro 200 is being reduced from 20× to 10×, while no published absolute quantity allows customers to verify what either 20× or 10× represents.

For EU customers, this may also raise a consumer transparency question because the quantity of a digital service can be an essential characteristic of the purchase.

How can customers verify what they are actually buying when neither the underlying allowance nor the calculation behind 5×, 10× or 25× is disclosed?

2

This is partly why I built Codexometer: track your quota, usage, orchestrate sessions & benchmark models!

We cannot know all the inner accounting but this project maintains a “best guess” at dollar API equivalent expenditure/capacity for your specific plan point.

See this part of the README.

3

Last couple of days have been exceptionally good with Codex max200 plan. It is giving me 4x the ROI on GPT 6.1 sol compared to what it give a week back. It could be possibly from the 95% discount on cached tokens, which play a significant role in long running tasks.

4

I’m on Pro 200 too, and this is what I’d like clarified before October 30. My email gives the date of the cut, but the percentage meter doesn’t tell me how much usage my account actually includes.

A tool like Codexometer can help estimate consumption. I’d still like OpenAI to show the allowance behind that percentage and let us see what each task used. If that allowance or the way usage is counted changes, put the date in the usage history. Then we could check the change ourselves instead of trying to work it out from how quickly the bar falls.