### Check for existing issues

- [x] I have searched the existing issues and checked that my issue is not a duplicate.

### What happened?

### What happened?

When selecting a team in the "Models + Endpoints" page (Current Team dropdown), the page shows **all model deployments** on the proxy instead of only the models belonging to the selected team's access group.

The API-level enforcement works correctly — `/v1/models` returns only the models the team has access to, and chat/embedding requests to unauthorized models are properly denied with 401. The issue is purely in the UI's model listing.

### Steps to reproduce

1. Configure two access groups with different models: - `Content-Platform` → gpt-4o, gpt-5.2, gpt-5.4, text-embedding-3-large, text-embedding-3-small - `Study-Partner` → gpt-5.4-pro, text-embedding-3-large, text-embedding-3-small

2. Create a team `Content Platform` with `models: ["Content-Platform"]`

3. Create a key scoped to that team

4. Log into the UI with that key

5. Go to **Models + Endpoints** → select **"Content Platform"** from the "Current Team" dropdown, with **View: "Current Team Models"**

6. **Expected:** Only the 5 models in the `Content-Platf…