🤖 refactor: centralize model constants + fix previousResponseId bug #584
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
- Fix GPT_MINI to use correct model ID (gpt-5.1-codex-mini) - Remove duplicated string constant exports (SONNET, GPT, etc.) - Callers now use KNOWN_MODELS.SONNET.id directly - Export KNOWN_MODELS object for direct access - Build MODEL_NAMES programmatically from KNOWN_MODELS - Groups by provider automatically - No manual duplication needed - Add integration test verifying all known models exist in models.json - Tests run outside IPC layer - Catches missing/renamed models early Generated with `mux`
Fixes APICallError when switching models or using non-reasoning models. Previously, we would pass previousResponseId from any previous OpenAI assistant message, causing errors when: - Switching between models (e.g., gpt-5.1-codex → gpt-5-pro) - Using models without reasoning support - Response IDs expired or invalid Now we: 1. Only extract previousResponseId when current model uses reasoning 2. Only use it if the previous message was from the same model 3. Stop searching if we encounter a different model (conversation context changed) This prevents "Previous response with id 'resp_...' not found" errors when the response ID is invalid for the current model/context. Generated with `mux`
- Add integration test for multi-turn conversations with reasoning models - Verifies previousResponseId is passed correctly between turns - Tests GPT_CODEX specifically to catch response ID bugs - Validates responseId exists in assistant message metadata - Update PROVIDER_CONFIGS to use KNOWN_MODELS constants - Ensures tests use the same model IDs as production code - Prevents drift between test models and known models Generated with `mux`
ammar-agent
force-pushed
the
centralize-model-constants
branch
from
c3807ea to
563e808
Compare
ammar-agent
force-pushed
the
centralize-model-constants
branch
from
563e808 to
8fb9ad1
Compare
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
ammar-agent
force-pushed
the
centralize-model-constants
branch
2 times, most recently
from
4d4a2a3 to
b141d20
Compare
ammar-agent
force-pushed
the
centralize-model-constants
branch
from
b141d20 to
7ca2a67
Compare
_Generated with `mux`_
The E2E test failure was caused by MODEL_ABBREVIATIONS being sorted alphabetically, which made gpt-5.1-codex the first model in the LRU instead of the default anthropic:claude-sonnet-4-5. Fix: Ensure defaultModel is always first in DEFAULT_MODELS array. _Generated with `mux`_
ammario
approved these changes
Nov 14, 2025
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
Centralizes all model metadata into
knownModels.tsand fixes a bug wherepreviousResponseIdwas incorrectly passed between different models or contexts.Changes
1. Centralized Model Constants (
knownModels.ts)GPT_MINImodel ID: Changed togpt-5.1-codex-mini(correct LiteLLM name)provider+providerModelIdKNOWN_MODELS.SONNET.iddirectlyMODEL_NAMES: Auto-groups models by provider via.reduce()KnownModelKeyderived from object keys, never out of syncmodels.json2. Fixed
previousResponseIdBugProblem: We were passing
previousResponseIdfrom any previous OpenAI message, causing:APICallError: Previous response with id 'resp_...' not foundwhen switching modelsSolution: Now only pass
previousResponseIdwhen:reasoningEffortis set)3. Test Improvements
PROVIDER_CONFIGSto useKNOWN_MODELSconstantsresponseIdpersistence across conversation turnsBenefits
Testing
Generated with
mux