The model 'gpt-image-2' does not exist.

Same

just got it right now, so annoying

+1

Mass outage?

same, I cannot use the app

+1 019e895d-7e87-7023-a732-e300d26a1910
(note my thread now works)

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

• #25966

Powered by Codex Action

+1 019e8a9e-fac9-7a01-b6bb-0c3a051c7ef3

Same issue! Maybe some backend change triggered an issue with that new update's image skills?

Tried changing models but no dice. Jeez.

+1 019e8aa8-af7b-7f10-8e9a-aeadd6f97fae

I went to skills and uninstalled the image generation skill as a temporary workaround.

• Model changed to gpt-5.4 medium

› test

• Received. I’m here and responding normally.

This works now.

Same 5.5 Extra High

+1

Okay it is back now for me, even with 5.5.

Adding a local Codex Desktop datapoint from macOS.

Environment:

• Codex Desktop: 26.601.20914 build 3497
• App-bundled CLI: codex-cli 0.136.0-alpha.2
• Global CLI: codex-cli 0.135.0
• macOS: 26.3 build 25D125, Darwin 25.3.0 arm64
• Thread model: gpt-5.5, reasoning xhigh

What I observed locally:

• The visible UI error card said: The model 'gpt-image-2' does not exist.
• The active work in the thread was Browser Use patch verification, not image generation.
• The literal error was not in the conversation rollout body for the browser-patch thread; it appeared in the desktop app log as a title-generation failure.

Relevant local log excerpt:

2026-06-02T23:24:29.420Z warning [electron-fetch-handler] Failed to generate thread title errorMessage="Structured turn failed: {
  "type": "error",
  "error": {
    "type": "image_generation_user_error",
    "code": "invalid_value",
    "message": "The model 'gpt-image-2' does not exist.",
    "param": "tools"
  },
  "status": 400
}"
2026-06-02T23:24:29.420Z error [electron-message-handler] Received turn/completed for unknown conversation conversationId=019e8aa7-111a-7751-bacf-ae1ebfabb6f7

The thread where I saw it:

• 019e8aa7-0ef8-7b81-9a99-9aa5b77392a5

Related browser-patch thread active at the time:

• 019e8a9c-bf36-7ad1-a419-d0deda26a44d

My read: this looks like an app/runtime-side hosted tool configuration problem on the structured turn used for title generation, not a failure caused by the user task or by a direct image generation request.

fixed.

The error just recurred for me.

+1

happened again.

Started happening again.

Sadly this has seriously disrupted our work, and we hope it can be fixed as soon as possible.

This problem has arisen again.

same issue, kinda stopped happening and then began again. no image generation promped, just using 5.5 high.

BS

Maybe fixed now?

+1

https://status.openai.com/incidents/01KT5B2V94YX1T1M88BVQHB58J

It keeps interrupting pipeline from time to time

This is happening in every thread I've tried from the codex app, new or old, since applying the most recent update. It recurs despite restarts and logging out and logging back in.

Same issue: The model 'gpt-image-2' does not exist. after any prompt while use codex with 5.4

Version 26.601.20914 • Released Jun 2, 2026

Workaround is to add this to your config.toml:

[features]
image_generation = false

Thanks for reporting the issue. This incident has been mitigated, so you should no longer see the error.