🤖 fix: route mux gateway models in backend#2025
Conversation
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 71c79a8440
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".
|
@codex review Addressed the legacy mux-gateway routing note by honoring explicit mux-gateway prefixes (with comments) while still respecting backend config + global toggle. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 51c28cc9f7
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".
|
@codex review Preserved explicit mux-gateway prefixes for legacy clients in send/resume before routing, while still normalizing persisted model IDs. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 52565615d8
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".
5256561 to
d3f6f04
Compare
Propagate routedThroughGateway metadata through stream events and message hydration so the UI can render the Mux Gateway indicator even when models remain canonical. --- _Generated with `mux` • Model: `openai:gpt-5.2-codex` • Thinking: `xhigh` • Cost: `$17.62`_ <!-- mux-attribution: model=openai:gpt-5.2-codex thinking=xhigh costs=17.62 -->
Codex noted that compaction metadata may contain an empty model string. Trim and fall back to workspace defaults to avoid backend validation errors. --- _Generated with `mux` • Model: `openai:gpt-5.2` • Thinking: `xhigh` • Cost: `$17.93`_ <!-- mux-attribution: model=openai:gpt-5.2 thinking=xhigh costs=17.93 -->
d3f6f04 to
07c935c
Compare
|
@codex review Fixed empty/whitespace compaction model overrides by trimming and falling back to the workspace model; added tests to cover both cases. |
|
Codex Review: Didn't find any major issues. More of your lovely PRs please. ℹ️ About Codex in GitHubCodex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
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". |
Add unit tests for AIService.resolveGatewayModelString to lock in routing behavior across enabled/disabled/configured cases, supported providers, variant allowlisting via modelKey, and explicit legacy opt-in. --- _Generated with `mux` • Model: `openai:gpt-5.2` • Thinking: `xhigh` • Cost: `$17.93`_ <!-- mux-attribution: model=openai:gpt-5.2 thinking=xhigh costs=17.93 -->
|
@codex review Follow-up: added unit tests for |
|
Codex Review: Didn't find any major issues. Chef's kiss. ℹ️ About Codex in GitHubCodex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
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". |
Summary
Background
Implementation
Validation
Risks
Generated with
mux• Model:openai:gpt-5.2-codex• Thinking:xhigh• Cost:$6.60