Codex responds too slowly

Open 💬 14 comments Opened May 25, 2026 by ty91
💡 Likely answer: A maintainer (github-actions[bot], contributor) responded on this thread — see the highlighted reply below.

What issue are you seeing?

I'm using Codex via the Codex CLI and also the Pi CLI. Responses have been coming too slowly since last Saturday, I think, especially when it falls back to SSE from WebSocket.

What steps can reproduce the bug?

Configure the Codex CLI via ~/.codex/config.toml to use SSE explicitly, like:

model_provider = "openai-sse"

[model_providers.openai-sse]
name = "OpenAI"
wire_api = "responses"
requires_openai_auth = true
supports_websockets = false

Then start a new session. Codex takes at least 5 minutes to answer.

What is the expected behavior?

_No response_

Additional information

_No response_

View original on GitHub ↗

14 Comments

GGBondBlueWhale · 3 months ago

I got the same issue in these few days.

github-actions[bot] contributor · 3 months ago

Potential duplicates detected. Please review them and close your issue if it is a duplicate.

  • #24045
  • #23665
  • #24419
  • #24414
  • #24422

Powered by Codex Action

ty91 · 3 months ago

@GGBondBlueWhale Yep, looks like there are a bunch of open issues already. I ran into this one while using Pi, and it seems the Codex backend's SSE connection has some problems.

Yorick-Ryu · 3 months ago

same issue

OverHash · 3 months ago

This appears to be linked to https://github.com/anomalyco/opencode/issues/29079. OpenAI's SSE endpoint appears to not respond well -- sometimes leaving an empty response, especially for oceania users.

adeperio · 3 months ago

Yep confirm that I'm seeing the issue on Opencode too

MisterRound · 3 months ago
DimaSergeew · 3 months ago

I encountered the same problem

jordan584 · 3 months ago

I got the same problem in codex app

jordan584 · 3 months ago

I can resolv this problem, Maybe you use vpn, and then you have five reconnecting, after that, is thinking. but it is too slow,
so you need to use TUN Mode clash. It can work to me

hannespreishuber · 3 months ago

unacceptable slow
generating an prompt for image generation takes minutes.
recreate a few kb of hmtl 15 minutes an up.
WIndows codex desktop app hangs several minutes.
Everyting ist 20x times faster in Chat Web Window. Payed 200$ for that shit.

icanall10 · 3 months ago

I use both Codex and Opencode with a chatgpt subscription. Codex is at least 3 times slower

wbdb · 2 months ago

@hannespreishuber 30 minutes one image :(, 20-30 minutes for small coding task.

AmberLee2427 · 2 months ago

Feedback ID: 019eb2d7-78e6-7bb2-94cd-702faaddf134