Codex CLI wont do anything
Resolved 💬 26 comments Opened Mar 9, 2026 by madziszyn Closed Mar 9, 2026
💡 Likely answer: A maintainer (github-actions[bot], contributor)
responded on this thread — see the highlighted reply below.
What version of Codex CLI is running?
codex --yolo
What subscription do you have?
Plus
Which model were you using?
gpt-5.4
What platform is your computer?
_No response_
What terminal emulator and version are you using (if applicable)?
_No response_
What issue are you seeing?
Codex wont do anything. wont respond (plan mode, normal mode)
What steps can reproduce the bug?
Uploaded thread: 019cd245-42cd-7471-8d7d-0e880d4c81d0
What is the expected behavior?
_No response_
Additional information
_No response_
26 Comments
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
Seems like codex currently has global issues with being veeery slow to respond. Same for me in codex app on mac. Many people in discord reported the same. Started about an hour ago.
i have a similar issue .. wondering if it is related. out of nowhere any input results in no response at all (not even "Working" .. i also cant ctrl-c out of the session. .. i keep starting new sessions, sometimes it works then in the next session it stops working again. maybe 1 out of 20 sessions it responds or starts working. I am a bit baffled
Very slow indeed , takes more than one minute to just reply "Hello".
How come https://status.openai.com/ says they are fully operationnal? Seems like they don't know about the issue in the first place
"We’re fully operational"
A bit annoying that there's obviously a major issue and after 45min there's zero acknowledgement anywhere. What's the purpose of status pages if it doesn't reflect current issues. ..
What discord server is this @Ivan-composer ?
@dayowe OAI official, codex channel https://discord.gg/openai
same reports on reddit https://www.reddit.com/r/codex/comments/1roxj50/codex_down_5min_wait_time_between_tool_calls/
dealing with the same issue, unbearably slow for simple ops
<img width="865" height="51" alt="Image" src="https://github.com/user-attachments/assets/86fd26ba-e8af-437d-8cc8-87b2e401c210" />
For me it seems back to normal rn
thanks @Ivan-composer
yeah same issue too, but after restart it works, but still slow and feels dumber
Same issue, not sure it is the API end, because if it starts working there are no issues.
The issue is randomly, codex doesn't respond to the initial command and CLI becomes unresponsive. After submitting the initial command, I see no "Working" indicator or able to quit using Ctrl-C. Seems like Codex just hangs.
Happened both with CLI and codex exec. In one occasion, Codex responded after a long wait.
For me the same.
When its not working the TUI hangs and i need to close it. It is working in like 25% of the requests for me.
same here for half an hour, looks like a major outage
Just downgrade the cli works way better
For me the same. had to switch to another model (GPT-5.3-Codex).
same for me. it was like this last week too at the same time.
Experiencing the same issues
Same issues here!
Same issue.
Same issue here I guess. But when I send a message it's not even loading anymore. Quitting does also not work. Seems completely dead. Is it the same for you?
Could be a EU centric problem, after many unsuccessful tries I switched to USA IP via VPN and it is working now.
Its so slow for first response! Getting unuseable.
After ~5 minutes of waiting, full access seems restored and responsive, This is a waiting time PER isolated instance (codex new thread, new cmd instance, etc!) Very strange, nothing is responsive for 5 minutes (Can't Ctrl+C out of the CLI, Can't stop the generation in the Codex app) and suddenly it works again.
This feels like a bug not being responsive to CTRL-C and no "working" indicator while being within a openai/server outrage.
@etraut-openai
still having the same issue. the answer come in very slow, we are talking about a single message/tool call every 5min
We're investigating the issue. I'm going to close this in favor of #14048 to consolidate. I'll post updates there as they become available.