### Check for existing issues

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

### What happened?

`GET /v1/models` and `GET /models` return `owned_by: "openai"` for every configured model, regardless of the model's actual backend provider. An Anthropic, Bedrock, Vertex AI, or Azure AI deployment shows up misreported as owned by OpenAI.

`model_list()` in `litellm/proxy/proxy_server.py` hardcodes `provider="openai"` at its `create_model_info_response()` callsites instead of resolving the real provider from the deployment. The sibling `GET /v1/models/{model_id}` endpoint already resolves it correctly, so only the plural listing endpoint is affected.

The response's `created` field is also always a fixed `1677610602` (March 2023) regardless of when the model was actually added, though that one is lower severity since it's filler most OpenAI-compatible clients ignore. This is also the subject of an issue that was previously closed as stale (https://github.com/BerriAI/litellm/issues/21371)

Fixed in #32901.

### Steps to Reproduce

1. Configure a proxy with a non-OpenAI model: ```yaml model_list: - model_name: claude-sonnet litellm_params…