improve the sync_model and model finder jobs

[Caitlin Pinn (cpinn)]

• modify the existing prompts to be more targeted both when finding and fixing particular issues
• add an llm step to the sync_models job to add a display name
• remove all embedding models. Embedding models do not work in the playground which will auto route to chat completions. Ensure no embedding models are pulling in through sync_model or model finder.

With embedding models I get the following error which makes sense since they are currently being routed to the chat/completions endpoint.
"message": "You are not allowed to sample from this model",

You can still call embeddings directly but they do not need to be part of the model registry.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
ai-proxy	Ready	Preview, Comment	Apr 29, 2026 5:36pm

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 9e1b240a8a

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 (@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 (@codex) address that feedback".

[Caitlin Pinn (cpinn)]

These were added during the large sync-everything change.

We only allow through chat and completions from the playground. We don't really have a view for embeddings at the moment, they need to be called directly.

Removing them from the list.

[aswink]

nice!