Usage limits normalized this morning but regressed by evening, draining ~5x faster again
Hi, opening a separate issue because this looks like a recurring/systematic problem, not a one-off spike.
Related previous issue: [https://github.com/openai/codex/issues/30939](<https://github.com/openai/codex/issues/30939>)
Previous Feedback ID: 019eaf5c-26e2-79a0-b678-ffa25afc0863
What changed today:
- In the morning of July 7, 2026, usage finally looked normal again.
- By the evening of July 7, 2026, the behavior regressed.
- Limits are again draining around 5x faster than expected.
This is happening even after I completed identity verification with passport verification. I am not posting any personal verification details here because this is a public GitHub issue, but support should be able to check the account internally.
This has now been recurring since around mid June. Earlier the limits were burning about 5-10x faster than normal, then it briefly improved, and now the same issue came back the same day.
Why this looks systemic:
- The issue comes and goes without a matching change in my actual Codex usage.
- It previously looked related to route/IP/anti-abuse filtering.
- A US connection made the behavior normal for a while, but that stopped helping.
- Today it was normal in the morning, then reverted again by evening.
- This is on a paid
100x5plan, so the product is effectively unreliable for normal work.
Environment:
- Codex CLI version:
codex-cli 0.142.0 - App: Codex Desktop
- OS: macOS
15.7.4(24G517) - Current region / connection: Poland, P4 Sp. z o.o. (
AS9141) - Plan: paid
100x5 - Started: around mid June 2026
- Regressed again: evening of July 7, 2026
Can you please investigate this as a recurring usage-accounting/rate-limit regression, check whether the account is still affected by anti-abuse/rate-limiter state, and reset the incorrectly drained quota for the affected period?
I can upload new feedback / screenshots if needed.
User feedback: unexpected usage drain in Linux app sessions
- Source: [https://x.com/bot_adeel/status/2090720124714312012](<https://x.com/bot_adeel/status/2090720124714312012>)
- Source: [https://x.com/madsholme/status/2090815302191755573](<https://x.com/madsholme/status/2090815302191755573>)
- Source: [https://x.com/bash_superishka/status/2091846329190240711](<https://x.com/bash_superishka/status/2091846329190240711>)
- Type: external
- User reports:
- “i'm on chatgpt app on linux, even for minor changes it consumes 3-4% of limit.”
- “I'm using the codex app on Linux and I'm severely degraded limits. Same with the cli agent.”
- “One 12min task on Sol Extra High took 5% of my weekly usage on the Plus plan.”
- “it wrote no code, it was just planning. 5% of usage for a single planning task on standard speed. I use the Chatgpt app on Ubuntu.”
Showing cached comments. Read the full discussion on GitHub ↗
9 Comments
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
Additional observation: in parallel with the limit-drain regression, I am also seeing noticeably slower inference / response generation. Sometimes I have to wait several minutes for a response, and subjectively the generation speed feels around ~0.5 tokens/sec.
This started showing up around the same time as the renewed fast limit drain, so it may be related to the same routing / rate-limiter / anti-abuse state rather than just usage accounting.
+1 Same observation for me max 5x limits draining 5x in mere about 4-5 messages. feedback id: 019f368b-a811-7343-9f3c-3184ef094dd0
Same here，i switched from pro 5x plan to pro 20x plan yesterday, and it seems 20x is drained as fast as 5x to me.
I've hit my 5h limits multiple mornings now on very light work using 5.4. I'm on the plus plan and usually never hit 5h limits. So, adding my voice, usage is totally messed up. The circled days are the two I have hit the 5hr cap and it is clearly nowhere near the usage of previous days (I only work 9-5, but it's very obvious even stacking 5x perfectly complete usage of every 5hr cap couldn't reach some of the other days usage).
<img width="1831" height="491" alt="Image" src="https://github.com/user-attachments/assets/c1aaa2b5-e499-4ade-b622-d434a7d1462f" />
It's also not because I'm using a high model as these days I was using lighter models predominantly because the tasks I ran were straightforward and not using big contexts.
<img width="1855" height="627" alt="Image" src="https://github.com/user-attachments/assets/ec4c38da-ba96-4b0f-ad26-27a23ca99ac1" />
Just the same here, my usage has gone down but I reach my limit even with the easiest prompts using GPT5.5 on medium
<img width="1080" height="1206" alt="Image" src="https://github.com/user-attachments/assets/a03d825e-4a6e-4214-b1f9-ab8008b6479b" />
These are the days I've hit 5hr usage. Literally my lowest usage working days in the month, and using lower quality models. How is this still not resolved or acknowledged?!
Some very strange things are going on, my usage keeps switching between:
<img width="1305" height="768" alt="Image" src="https://github.com/user-attachments/assets/473d30a8-5081-4425-a872-550db2d8f3b5" />
and:
<img width="1263" height="759" alt="Image" src="https://github.com/user-attachments/assets/adfe84ae-9685-472b-983e-1a6add356741" />
Same problem for me its insane Feedback ID: 019f53b9-e40d-7930-9014-7fea72713e7a