openai / openai/codex

[VS Code] 404 Unknown error after stopping a task, switching to GPT-5.6 SOL Ultra, and retrying

Open
#42,550 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

bug connectivity extension windows-os
Dominant language
Rust
Stars
125k
Forks
19.4k
PR merge metrics
PR metrics pending

Description

What version of the IDE extension are you using?

Codex - 26.825.51511

What subscription do you have?

pro

Which IDE are you using?

vscode

What platform is your computer?

Windows 11, x64 (build 26200.9168)

What issue are you seeing?

After stopping a running Codex task because I noticed that it was using a lower model/reasoning setting than I intended, I switched the model to GPT-5.6 SOL with High reasoning and retried the task.

After doing this, the conversation repeatedly enters:

Reconnecting 1/5
Reconnecting 2/5
...
Reconnecting 5/5

and then fails with:

unexpected status 404 Not Found: Unknown error, url: https://chatgpt.com/backend-api/codex/responses

Retrying the request in the same conversation produces the same reconnect sequence and the same 404 error, so I can no longer continue the task normally.

The issue appeared immediately after stopping the previous run, changing the model/reasoning setting, and retrying.

CF-Ray: [a355a47d584ba422-ICN]

What steps can reproduce the bug?
  1. Open a repository in VS Code and open the Codex IDE extension.

  2. Start a task using a lower model or reasoning setting than intended.

  3. Stop/cancel the running task before it completes.

  4. Change the model to GPT-5.6 SOL and set the reasoning level to High.

  5. Retry the task in the same conversation/thread.

  6. Codex** starts reconnecting repeatedly from Reconnecting 1/5 through Reconnecting 5/5.

  7. After the final reconnect attempt, the request fails with:
    unexpected status 404 Not Found: Unknown error, url: https://chatgpt.com/backend-api/codex/responses

  8. Retrying again in the same thread results in the same behavior.

What is the expected behavior?

After stopping a running task and changing the model, Codex should start a new request using the newly selected model and reasoning setting.

The existing conversation should remain usable without entering a repeated reconnect loop.

If the existing session cannot be reused after a model change, Codex should reset or recreate the session automatically, or display an actionable error message instead of retrying five times and returning an HTTP 404.

Additional information

No response

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Reproduce the stop, model change, and retry sequence in the VS Code extension, then trace its reconnect handling for the /backend-api/codex/responses endpoint. Check how the existing conversation is reused after cancellation and model changes. Done means the retried task remains usable or produces an actionable error without the repeated five-attempt 404 loop.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust, vscode
Domain
api, devtools
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
55/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.