fix: 'max' effort no longer 400s on pre-5.6 Codex Responses models (#68365 verified live) - #90441
Merged
Merged
Conversation
…longer 400s on 'max' (#68365 confirmed live) Live probes against api.openai.com/v1/responses (Aug 2026): - gpt-5.6: accepts none/low/medium/high/xhigh/max; rejects minimal, ultra - gpt-5.5: accepts none/low/medium/high/xhigh; rejects max ('Unsupported value'), minimal, ultra So #68365's premise was half right: 'max' does 400 — but only on pre-5.6 models; blanket-clamping max->xhigh on gpt-5.6 (its fix) would have capped the one model that supports max. The declared-vocabulary design absorbs this as data: codex_supported_efforts(model) picks CODEX_GPT56_EFFORTS or CODEX_LEGACY_EFFORTS, and the shared clamp does the rest. Both the main Codex transport and the auxiliary client's Responses path use it. Wire outcomes: ultra -> max on gpt-5.6, ultra/max -> xhigh on gpt-5.5/o5, minimal -> low everywhere.
2 tasks
Contributor
૮ >ﻌ< ა ci reviewran on 1aaa3d4 — chore: retrigger CI (zero-job dispatch failure, auto-heal)
|
alt-glitch
added
type/bug
This was referenced Aug 20, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
reasoning_effort: maxno longer 400s on pre-5.6 OpenAI Responses models — the Codex wire vocabulary is now per-model, verified by live probes (#68365's premise confirmed).Live results against
api.openai.com/v1/responses(Aug 2026):So @joelbrilliant's #68365 was half right:
maxdoes 400 — but only on pre-5.6 models. Its blanketmax → xhighclamp would have capped gpt-5.6, the one model that supportsmax. The declared-vocabulary architecture (#90350) absorbs the truth as data instead.Changes
agent/reasoning_effort.py:CODEX_GPT56_EFFORTS/CODEX_LEGACY_EFFORTS+codex_supported_efforts(model)(oldCODEX_RESPONSES_EFFORTSkept as alias)agent/transports/codex.py+agent/auxiliary_client.py: both Responses paths pick the per-model set; the shared clamp does the restWire outcomes:
ultra → maxon gpt-5.6;ultra/max → xhighon gpt-5.5/o5;minimal → loweverywhere.Validation
509 targeted tests green (reasoning-effort module, both transports, aux client). New test pins the per-model sets and the consequential clamps, quoting the live error strings.
Resolves #68365 (premise verified live; credit to @joelbrilliant for the report and observation).
Infographic