Possible quota/accounting issue: Pro weekly limit depletes unusually fast

What version of the Codex App are you using (From “About Codex” dialog)?

26.318.11754 (1100)

What subscription do you have?

ChatGPT Pro

What platform is your computer?

Darwin 25.3.0 arm64 arm

What issue are you seeing?

I’m seeing what looks like unusually aggressive weekly request limit consumption on my ChatGPT Pro plan when using GPT-5.4 xhigh through Codex.

I upgraded from Plus to Pro about 2 months ago because Plus was not sufficient for my usage. Until recently, the experience was acceptable, but over approximately the last week, my weekly remaining request limit has been depleting much faster than expected.

This is happening while using both:

• VSCode Codex extension
• macOS Codex App

The main issue is that the remaining weekly quota seems to drop disproportionately fast relative to my actual usage patterns. I do use GPT-5.4 xhigh heavily, but the current rate of weekly limit depletion feels materially different from earlier weeks and much more aggressive than expected.

I will attach both weekly and monthly usage screenshots to this issue. The screenshots show the recent usage pattern across Extension and Desktop App, and the recent depletion behavior is noticeably different from what I was previously experiencing.

What steps can reproduce the bug?

1. Be on a ChatGPT Pro subscription.
2. Use GPT-5.4 xhigh through the VSCode Codex extension and/or macOS Codex App.
3. Perform normal coding workflows across, such as:

• asking for code explanations
• refactoring code
• debugging
• follow-up implementation prompts
• iterative multi-turn coding tasks

1. Monitor the weekly remaining request limit in the usage dashboard.
2. Compare the depletion rate against prior weeks of similar usage.

Observed behavior:
The weekly remaining request limit decreases much faster than expected. This became especially noticeable over the last week.

Additional reproduction notes:

• This does not appear tied to one specific prompt or one single session.
• I observe similar behavior across both the VSCode Codex extension and the macOS Codex App.
• There is no clear error message; the issue appears to be related to usage/quota accounting or metering.
• I will attach weekly and monthly usage screenshots for reference.

What is the expected behavior?

I expect weekly request limit consumption to be consistent, predictable, and proportional to actual usage.

Even if GPT-5.4 xhigh has a higher usage cost than other modes, the current depletion rate feels abnormally high compared with my earlier Pro usage.

I would expect:

• similar usage patterns to produce roughly similar weekly depletion,
• consistent metering across Codex clients,
• and clearer visibility if GPT-5.4 xhigh is weighted significantly differently in quota accounting.

If this behavior is intended, it would help to have clearer transparency around how requests are counted and whether Codex clients trigger additional usage beyond what is obvious from the user side.

Additional information

<img width="1237" height="425" alt="Image" src="https://github.com/user-attachments/assets/c8ae42cf-3517-4f15-afdc-e177c2839c8c" />

<img width="1237" height="425" alt="Image" src="https://github.com/user-attachments/assets/d5e5dc10-2b2c-4ba5-91bb-588a49b4d1b6" />