Conversation
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Problem
f1d6988 refreshed
codex_client_models.json, movingmax_context_windowfor the GPT-5.6 family from 921000 to 872000 and addinggpt-6-astraat 872000. The matchingcontext_lengthvalues in thecodex-prosection ofmodels.jsonwere not moved, so the two files now disagree:models.jsoncodex-procodex_client_models.jsonmax_context_windowgpt-6-astragpt-5.6-solgpt-5.6-terragpt-5.6-lunaAstra is the odd one out in a second way. Its 272000 is the
context_windowvalue, while its three neighbours usemax_context_windowsemantics, so Astra currently advertises a smaller context than Sol despite being the higher-priority model.Precedent
34a122f, "Update GPT-5.6 context length to 921000 for Codex Pro and client models", changed
context_lengthinmodels.jsonandmax_context_windowincodex_client_models.jsonin the same commit, to the same number. This PR applies that same pairing to the value f1d6988 just landed.Evidence
GET https://chatgpt.com/backend-api/codex/models?client_version=0.153.2on a Pro account, captured under both originators:context_windowmax_context_windowgpt-6-astragpt-5.6-solgpt-5.6-terragpt-5.6-lunagpt-5.5codex-tuiandcodex_cli_rsagree on every row. gpt-5.5 is included as a control: it reports 272000 for both fields and is already stored as 272000, so it needs no change and is left untouched.Scope
Four
context_lengthvalues in thecodex-prosection, set to 872000. Nothing else is modified.codex-teamandcodex-plusstill carry 372000 for the 5.6 family and 272000 for Astra. I have left those alone deliberately: I only have a Pro credential, so I cannot observe what those plans actually return, and I would rather not change a number I cannot verify. If 372000 is a real per-plan limit then Astra's 272000 in those two sections is probably also wrong, but that needs someone on a Plus or Team account to confirm.