Model picker should show provider instance name for duplicate Ollama models

[fishcharlie]

• Copilot Chat Extension Version: Unknown (reported on Apr 22, 2026)
• VS Code Version: Code - OSS dev build from main
• OS Version: macOS (reporter) and Linux (second Ollama host)
• Feature (e.g. agent/edit/ask mode): Chat model picker
• Selected model (e.g. GPT 4.1, Claude 3.7 Sonnet): Ollama models (duplicate model ids across two providers)
• Logs: N/A

Steps to Reproduce:

1. Configure two Ollama providers in Copilot BYOK with different names (for example "Local MacBook" and "Remote Linux").
2. Ensure both providers expose at least one model with the same model id/name.
3. Open Chat and open the model picker.
4. Observe both duplicate entries show the generic label "Ollama" instead of the user-configured provider instance name.

Expected:
Duplicate models should show the configured provider instance/group name so users can distinguish which backend they are selecting.

Actual:
Duplicate models only show the generic provider label ("Ollama"), making it unclear which configured provider instance is selected.

Motivation:
I use two Ollama providers (one on a MacBook Pro, one on a Linux server). I need to pick the correct backend when model names overlap.

Proposed Fix:
In grouped language-model resolution, set model detail text to the configured provider group name so model picker rows display instance-specific labels (for example "Local MacBook" vs "Remote Linux").