fix(codex): correct stale-timeout hint — gpt-5.4-codex doesn't exist#33128
Closed
teknium1 wants to merge 1 commit into
Closed
fix(codex): correct stale-timeout hint — gpt-5.4-codex doesn't exist#33128teknium1 wants to merge 1 commit into
teknium1 wants to merge 1 commit into
Conversation
… soften framing The Codex silent-hang hint suggested `gpt-5.4-codex` as the workaround, but that model name doesn't exist in the OpenAI Codex catalog (the Codex-tagged SKUs are gpt-5.3-codex and gpt-5.2-codex; plain gpt-5.4 is the actual sibling model). Users following the hint hit a second error trying to switch to a phantom model. Also softens the framing from "appears to be silently rejecting" to "did not return any stream events ... within the stale-call timeout". The original phrasing asserted backend rejection as fact, but the detector only knows about the timeout — it can't actually tell whether the backend rejected the request, the network dropped, or latency spiked. Misleading users into thinking gpt-5.5 is broken when it's often transient. Changes - run_agent.py: rewrite the user-facing message — drop the false "silently rejecting" claim, point to gpt-5.4 (which exists), suggest retry first. - tests/run_agent/test_codex_silent_hang_hint.py: update assertions to match the new message; rename the negative-case test to use gpt-5.4 (the actual workaround) instead of the bogus gpt-5.4-codex. Validation - tests/run_agent/test_codex_silent_hang_hint.py → 10/10 passing
Contributor
🔎 Lint report:
|
Contributor
Author
|
Closing in favor of PR #33034 (@EvilHumphrey), which is a strict improvement. Their version:
Salvaging #33034 onto current main now. Sorry for the wasted CI cycle, missed their PR in the duplicate sweep. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
The Codex silent-hang hint suggested
gpt-5.4-codexas the workaround whengpt-5.5stalls on the chatgpt.com backend, but that model name doesn't exist in the OpenAI Codex catalog. The Codex-tagged SKUs aregpt-5.3-codex/gpt-5.2-codex; the plain sibling that works isgpt-5.4. Users following the hint hit a second error trying to switch to a phantom model.Also softens the framing: the previous text asserted "appears to be silently rejecting" as fact, but the detector only knows the stale-timeout fired — it can't actually distinguish backend rejection from network drop or latency spike. The original phrasing led users to believe
gpt-5.5was broken when it's often transient.Changes
run_agent.py: rewrite the user-facing message — drop the false "silently rejecting" claim, point togpt-5.4(which exists), suggest retry first.tests/run_agent/test_codex_silent_hang_hint.py: update assertions; rename negative-case test fromgpt-5.4-codex(phantom) togpt-5.4(real).Validation
gpt-5.4-codex(doesn't exist)gpt-5.4(real)User report that surfaced this: ChatGPT Plus account getting the stale-timeout hint, following the suggestion, hitting another error because
gpt-5.4-codexis not a known model.