### Check for existing issues

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

### The Feature

Add Modelsell as a registered OpenAI-compatible provider in LiteLLM.

Modelsell exposes:

- API base: `https://modelsell.com/v1` - Chat Completions: `POST /v1/chat/completions` - Model discovery: `GET /v1/models` - Authentication: Bearer API key - API key page: https://modelsell.com/console/token - Documentation: https://modelsell.com/docs/api-reference - Website: https://modelsell.com

A focused first implementation could use LiteLLM's existing JSON-configured OpenAI-compatible provider registration pattern, with Modelsell's base URL and authentication defaults. The model catalog should remain dynamically discoverable through `/v1/models` rather than being hardcoded.

Suggested initial scope:

1. Register `modelsell` as an OpenAI-compatible provider. 2. Support provider selection in the SDK and Proxy using the existing custom OpenAI routing path. 3. Add configuration documentation and focused routing tests. 4. Treat static model pricing metadata as a follow-up unless the maintainers prefer it in the first contribution.

I can contribute the imple…