Show which model answered in the ChatMessage footer

[ghostiee-11]

With per-task model_kwargs and the routing work in #2015, the model behind a reply is no longer fixed, but nothing in the UI says which one ran. Putting it in the message footer would make that visible.

The footer already carries the suggestion and retry buttons in lumen/ai/ui.py, so the slot exists. Two gaps:

• The resolved model name never reaches the caller. Llm._get_model_kwargs resolves it inside get_client, and invoke/stream return only the response.
• panel_material_ui.ChatMessage renders footer_objects below the edit/copy/reaction icons rather than under the bubble, filed upstream as ChatMessage renders footer_objects below the icon row and lays header/footer out in a row panel-extensions/panel-material-ui#719.

Raised in #2015 (comment)