v1.59.1.0 codex: refresh model docs to GPT-5.6 family (+ max reasoning level)#2213
Open
garrytan wants to merge 7 commits into
Open
v1.59.1.0 codex: refresh model docs to GPT-5.6 family (+ max reasoning level)#2213garrytan wants to merge 7 commits into
garrytan wants to merge 7 commits into
Conversation
Keeps the no-hardcode default (auto-frontier). Updates the illustrative model names in the Model & Reasoning section to the current GPT-5.6 family (sol/terra/luna, bare gpt-5.6 -> sol) and documents the new `max` reasoning-effort level above xhigh.
The branch bumped VERSION without package.json; test/gen-skill-docs.test.ts:256 asserts they match, so bun test failed on the branch as submitted. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…-5.6 alias
Live probes against the OpenAI API (2026-07-09):
- -c model_reasoning_effort=max works only on codex CLI >= 0.144 AND an explicit
GPT-5.6 model (-m gpt-5.6-sol). Installed 0.136 rejects the config value
('unknown variant max'); non-5.6 default models return API 400.
- Bare '-m gpt-5.6' is rejected for Codex with ChatGPT sign-in even though the
API-level alias to gpt-5.6-sol is real (OpenAI API changelog 2026-07-09).
- An 'ultra' tier exists above max on some surfaces (ChatGPT model picker,
codex-rs ReasoningEffort enum).
Template + regenerated SKILL.md in one commit (check-freshness bisect invariant).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…r credit, voice - gpt-5.2 is deprecated but still selectable (survived the 2026-04-14 Codex retirement wave; no API shutdown for the base model) — only gpt-5.1-codex-max is retired (API shutdown 2026-07-23). Entry no longer calls both 'retired'. - max escape hatch now carries the codex >= 0.144 + explicit-5.6-model caveats; notes the ultra tier above max. - Adds 'Contributed by @cathrynlavery' per repo convention. - Release-summary prose aligned with voice rules (no em dashes). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Filed from the PR #2211 full-checks run: fork shards burned 31m48s timing out on an empty ANTHROPIC_API_KEY before failing, and the contributor re-triggered the doomed run reading it as flakiness. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
|
Merging to
After your PR is submitted to the merge queue, this comment will be automatically updated with its status. If the PR fails, failure details will also be posted here |
E2E Evals: ✅ PASS2/2 tests passed | $1.16 total cost | 13 parallel runners
13x ubicloud-standard-8 (Docker: pre-baked toolchain + deps) | wall clock ≈ slowest suite |
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.
Continues #2211 by @cathrynlavery from a base-repo branch. Fork PRs don't receive the repo's eval CI secrets, so the eval shards on #2211 could never authenticate (empty
ANTHROPIC_API_KEY→error_api/ login timeouts). Her commits are preserved with her authorship; four follow-up commits from us ride on top.What this ships (her work)
Documentation-only refresh of
/codexfor the GPT-5.6 model family (sol>terra>luna):-mexamples updated fromgpt-5.1-codex-max/gpt-5.2, Model & Reasoning section names the current frontier, and the newmaxreasoning level is documented. No behavior change —/codexstill hardcodes no model. Contributed by @cathrynlavery.Follow-up commits (ours)
test/gen-skill-docs.test.ts:256asserts they match, sobun testfailed as submitted.-c model_reasoning_effort=maxneeds codex CLI >= 0.144 AND an explicit 5.6 model (-m gpt-5.6-sol→ works; default model → API 400; CLI 0.136 rejects the value at config parse).-m gpt-5.6is rejected for Codex with ChatGPT sign-in even though the API-level alias tosolis real (OpenAI API changelog 2026-07-09).ultratier abovemaxon some surfaces.gpt-5.2is deprecated-but-selectable, not "retired" (onlygpt-5.1-codex-maxis retired, API shutdown 2026-07-23); adds the contributor credit line; release-summary voice cleanup.Verification
-m gpt-5.6-sol/-terra/-lunaall work;=maxworks on 0.144.1 with-m gpt-5.6-sol.eval:select): 46 pass / 2 fail / 324 skip. Both failures reproduce identically on clean origin/main (autoplan-dual-voice Phase-1 claude-voice assertion; gemini live adapter fails on local gemini CLI auth:IneligibleTierError). Pre-existing with receipts, not this diff.gen:skill-docszero drift; CHANGELOG ordering clean; VERSION 1.59.1.0 confirmed against the open-PR queue (queue-advanced past v1.59.0.0 feat: /land skill + trunk.io merge-queue integration #1815's claimed 1.59.0.0).🤖 Generated with Claude Code