Codex usage draining fast? Wrong plan limits, now fixed

Codex usage draining fast? 40,000 accounts had the wrong plan limits

Updated 2026-10-20. From Tibo's Oct 20 post and OpenAI's Codex pricing page.

Quick answer

On Oct 20, 2026 at 1:10 PM ET, Tibo (@thsottiaux) at OpenAI said about 40,000 accounts had usage limits over the weekend that did not match their selected plan, so usage drained faster than expected. OpenAI says the issue is now corrected, it is contacting each affected user, and affected accounts will get credits. This is not a reset for everyone. Check your current limits at chatgpt.com/codex/settings/usage or with /status in the Codex CLI.

What we know and what we don't

Question	Answer
How many accounts	About 40,000
What went wrong	The usage limits applied did not match the plan the account had selected
When	Over the weekend before Oct 20 (Oct 17 and 18)
Is it fixed?	Yes, OpenAI says it corrected the issue
What affected users get	Credits added to their account
How you find out	OpenAI says it is reaching out to each affected user
Which plans were hit	Not said
How many credits	Not said
When credits arrive	Not said
A reset for everyone?	No. This post is a fix plus credits, not a global reset

From Tibo's post on Oct 20, 2026 at 1:10 PM ET. Checked 2026-10-20; we watch @thsottiaux every hour and update this page if OpenAI adds detail.

What OpenAI said

"There were reports of usage draining faster than anticipated for some users over the weekend. We've identified approximately 40,000 accounts that had an inconsistency on which usage limits where the actual usage that was applied did not correspond to their actual selected plan. We are reaching out individually to all these users and have corrected the issue. If this affected you, we'll add credits to your account. Sorry we got this wrong."

Tibo (@thsottiaux), who works on Codex at OpenAI, on X, Oct 20, 2026, 1:10 PM ET. The post does not name a product, but Codex and ChatGPT Work share one usage allowance, and most of the weekend complaints came from Codex users.

No follow-up from Tibo or OpenAI on this yet (checked 2026-10-20).

How to check your Codex usage now

Usage dashboard: open chatgpt.com/codex/settings/usage. It shows your current limits and reset times.
In the CLI: type /status in an active Codex session to see what is left.
In the app: Settings > Usage in the Codex desktop app shows the same numbers, plus any banked reset you can apply.
Check the account: make sure you are signed in to the account or workspace you pay for. A personal login and a Business workspace have separate limits.

For scale, OpenAI's own estimate for Plus is "Estimated 15 to 160 local GPT-6.1 Sol messages per 5 hours; weekly limits may apply; shared with ChatGPT Work" (Codex pricing page, checked 2026-10-20). Pro plans currently have no 5-hour limit, only weekly limits. These are estimates, not fixed caps, so a single heavy day is not proof you were one of the 40,000.

If you think you were affected

OpenAI says it is reaching out to every affected account and will add credits. You should not need to file a ticket to qualify.
The fix stops the wrong limits from here on. It does not refill what already drained; the credits are the make-good.
If you are out right now: Plus and Pro can buy extra credits, anyone can run local chats with an API key at API rates, or you can wait for your window to roll over.

Other reasons Codex usage drains fast

If you were not one of the 40,000 and usage still vanishes, these are the usual causes, per OpenAI's pricing page:

Speed modes: Fast uses included limits at 2.5x and Ultrafast at 8x the Standard rate.
Model choice: on Plus, OpenAI estimates 5 to 45 local GPT-6 Astra messages per 5 hours vs 15 to 160 on GPT-6.1 Sol. See GPT-6 Astra vs GPT-6.1 Sol.
Images: image generation uses included limits about 3 to 5x faster than similar turns.
Shared pool: local messages, cloud tasks and ChatGPT Work draw from one allowance (and ChatGPT for Excel on Plus and Pro).
Context: large AGENTS.md files and many MCP servers add tokens to every message.

Was there a reset?

Not with this post. The last real Codex reset was Day 15 of the 28-day run: limit reset for Codex and ChatGPT Work on Oct 20 at 2:33 AM ET (post), before the fix was announced. Upcoming and past resets are on when does Codex reset and the live resets page, and every Tibo reset post is in the Codex reset log.

Sources

FAQ

Why is my Codex usage draining so fast?

OpenAI says about 40,000 accounts had usage limits over the weekend of Oct 17 and 18 that did not match their selected plan. That is now fixed. Other common causes are Fast or Ultrafast mode, GPT-6 Astra, image generation and large context.

Was Codex using the wrong plan's limits?

For about 40,000 accounts, yes. Tibo (@thsottiaux) said on Oct 20, 2026 that the usage limits applied to them did not match their selected plan.

Will I get credits back?

If your account was affected, OpenAI says it will add credits and is reaching out to each affected user. It has not said how many credits or when.

How do I check my Codex usage?

Open chatgpt.com/codex/settings/usage, type /status in an active Codex CLI session, or go to Settings > Usage in the desktop app.

Did OpenAI reset Codex limits for everyone?

Not with this announcement. It was a fix for the affected accounts plus credits. Global resets are tracked on the LimitDigest resets page.

Is it fixed?

OpenAI says it has corrected the issue, so limits should now match your plan.