fix: add gpt-5.3-codex-spark to openai-codex curated model list by teknium1 · Pull Request #7844 · NousResearch/hermes-agent

teknium1 · 2026-04-11T16:59:25Z

Summary

_PROVIDER_MODELS['openai-codex'] was missing gpt-5.3-codex-spark, causing it to not appear in the gateway /model picker (Telegram/Discord) for ChatGPT OAuth users.

The CLI hermes model flow gets codex-spark via codex_models.py's _add_forward_compat_models() synthesis — but the gateway reads the static curated list from models.py directly, which was never updated.

Follows up on #7670 which added gpt-5.4 and gpt-5.4-mini but missed this one.

Changes

Added gpt-5.3-codex-spark to _PROVIDER_MODELS['openai-codex'] in hermes_cli/models.py

Test plan

Gateway /model picker for openai-codex users should now show codex-spark in the list

The _PROVIDER_MODELS['openai-codex'] static list was a manually maintained duplicate of DEFAULT_CODEX_MODELS in codex_models.py. They drifted — the static list was missing gpt-5.3-codex-spark (and previously gpt-5.4). Replace the hardcoded list with _codex_curated_models() which calls DEFAULT_CODEX_MODELS + _add_forward_compat_models() from codex_models.py. Now both the CLI 'hermes model' flow and the gateway /model picker derive from the same source of truth. New models added to DEFAULT_CODEX_MODELS or _FORWARD_COMPAT_TEMPLATE_MODELS automatically appear everywhere.

…trol backoff Merge upstream/main (3 commits): - fix(streaming): adaptive backoff + cursor strip to prevent message truncation (NousResearch#7683) Telegram flood control during streaming caused messages to be cut off. New: adaptive backoff (double interval on flood, cap at 10s), cursor strip on fallback entry, retry on fallback send, default edit_interval 1.0s. - fix: unify openai-codex model list from codex_models.py (NousResearch#7844) - fix(matrix): pass required args to MemoryCryptoStore for mautrix >=0.21 (NousResearch#7848) Add 33 tests for the new streaming flood control behavior: - TestIsFloodError (7): flood/retry/rate detection, false positives - TestSendOrEditReturnValues (9): bool return values for all scenarios - TestAdaptiveBackoffDirect (6): progressive backoff, strike tracking, reset - TestTryStripCursor (4): cursor cleanup on fallback entry - TestFallbackFinalRetry (1): retry on flood during fallback send - TestDefaultEditInterval (3): config defaults - TestFloodStrrikesState (3): initialization state Self-improve: automated improvement

…sResearch#7844) The _PROVIDER_MODELS['openai-codex'] static list was a manually maintained duplicate of DEFAULT_CODEX_MODELS in codex_models.py. They drifted — the static list was missing gpt-5.3-codex-spark (and previously gpt-5.4). Replace the hardcoded list with _codex_curated_models() which calls DEFAULT_CODEX_MODELS + _add_forward_compat_models() from codex_models.py. Now both the CLI 'hermes model' flow and the gateway /model picker derive from the same source of truth. New models added to DEFAULT_CODEX_MODELS or _FORWARD_COMPAT_TEMPLATE_MODELS automatically appear everywhere.

Follow-up to the static list refresh: replace the hardcoded xAI entries with _xai_curated_models(), mirroring the _codex_curated_models() pattern from PR #7844. The helper reads $HERMES_HOME/models_dev_cache.json at import time (no network call) and falls back to a small static list when the cache is missing or malformed. Why: _PROVIDER_MODELS["xai"] has drifted once already (issue #16699) and will drift again next time xAI renames a model. Hermes already maintains the models.dev cache and uses it for context-length lookups; pointing _PROVIDER_MODELS at the same source means the /model picker self-heals on the next cache refresh instead of requiring a PR. Behavior: - With cache populated (normal user): shows every current xAI model ID, picks up renames automatically on next refresh. - Without cache (fresh install, offline): falls back to a static snapshot of the 9 current flagship IDs. - Malformed cache / unexpected shape: same static fallback, no crash. Import time verified <20ms — disk read only, no HTTP. Addresses the structural piece of #16699 ("consider a single _provider_models(provider) resolver") for xAI. Other per-provider lists can adopt the same pattern as drift is observed.

…sResearch#7844) The _PROVIDER_MODELS['openai-codex'] static list was a manually maintained duplicate of DEFAULT_CODEX_MODELS in codex_models.py. They drifted — the static list was missing gpt-5.3-codex-spark (and previously gpt-5.4). Replace the hardcoded list with _codex_curated_models() which calls DEFAULT_CODEX_MODELS + _add_forward_compat_models() from codex_models.py. Now both the CLI 'hermes model' flow and the gateway /model picker derive from the same source of truth. New models added to DEFAULT_CODEX_MODELS or _FORWARD_COMPAT_TEMPLATE_MODELS automatically appear everywhere.

…search#16699) Follow-up to the static list refresh: replace the hardcoded xAI entries with _xai_curated_models(), mirroring the _codex_curated_models() pattern from PR NousResearch#7844. The helper reads $HERMES_HOME/models_dev_cache.json at import time (no network call) and falls back to a small static list when the cache is missing or malformed. Why: _PROVIDER_MODELS["xai"] has drifted once already (issue NousResearch#16699) and will drift again next time xAI renames a model. Hermes already maintains the models.dev cache and uses it for context-length lookups; pointing _PROVIDER_MODELS at the same source means the /model picker self-heals on the next cache refresh instead of requiring a PR. Behavior: - With cache populated (normal user): shows every current xAI model ID, picks up renames automatically on next refresh. - Without cache (fresh install, offline): falls back to a static snapshot of the 9 current flagship IDs. - Malformed cache / unexpected shape: same static fallback, no crash. Import time verified <20ms — disk read only, no HTTP. Addresses the structural piece of NousResearch#16699 ("consider a single _provider_models(provider) resolver") for xAI. Other per-provider lists can adopt the same pattern as drift is observed.