fix(models): add gpt-5.4 and gpt-5.4-mini to openai-codex static model list

[ygd58]

Summary

Fixes #6595

Root Cause

_PROVIDER_MODELS["openai-codex"] was stale (v0.7.0 era) and missing the current production models gpt-5.4 and gpt-5.4-mini.

provider_model_ids() already calls get_codex_model_ids() for dynamic discovery, but when the Codex API is unreachable the fallback static list was shown in the /model picker — without gpt-5.4.

Fix

Prepend gpt-5.4 and gpt-5.4-mini to the static fallback list so they appear in the picker even when the live API is unavailable.

Changes

• hermes_cli/models.py: prepend gpt-5.4 and gpt-5.4-mini to _PROVIDER_MODELS["openai-codex"]

[teknium1]

Thanks for the contribution @ygd58! This fix was already applied to main by an earlier commit.

Automated hermes-sweeper review — closing as implemented on main.

• gpt-5.4 and gpt-5.4-mini were added to the openai-codex model list in commit 671d5068e (PR fix: add gpt-5.4 to openai-codex model picker #7670).
• Subsequently, commit c94936839 (PR fix: add gpt-5.3-codex-spark to openai-codex curated model list #7844) refactored _PROVIDER_MODELS["openai-codex"] entirely: instead of a static list, it now calls _codex_curated_models() which derives models from DEFAULT_CODEX_MODELS in hermes_cli/codex_models.py — where both models appear at lines 15–16.
• The static list this PR would patch (_PROVIDER_MODELS["openai-codex"]) no longer exists; the approach is superseded.

No further changes are needed here.