Optional: Compare models
Time: 10 minutes Prerequisite: Complete the seven core steps first.
What you’ll compare
Section titled “What you’ll compare”Step 1 introduced ModelSelector, which lists the models available to your signed-in account and
stores the selected model ID in selectedModel. Every session configuration in the workshop uses
that value through Model = selectedModel.
Try a different model
Section titled “Try a different model”Run the completed application again:
dotnet run --project workshop-appChoose a different number when the picker lists available models, then enter the workshop target URL. The application architecture, local tools, MCP configuration, and permission policy remain the same; only the model assigned to the session changes.
What selection changes
Section titled “What selection changes”Model choice can affect latency, capability, and billing. It does not grant new tools or change the browser permission boundary. Keep tool allowlists and permission handling identical while you compare the response.
Troubleshooting this extension
| Symptom | Fix |
|---|---|
| No models are listed | The helper falls back to the account default; verify authentication if this is unexpected. |
| A number is outside the range | The helper safely uses the first model. |
| Tools disappear | Keep Model = selectedModel alongside the existing Tools, McpServers, and OnPermissionRequest settings. |
The extension is complete when: you can name the selected model and explain why the same scoped tools and permissions still apply.
Check your understanding
Section titled “Check your understanding”What does choosing a model change, and what does it leave unchanged?
Check your answer
It selects the model used by a CopilotSession, which can affect the response and latency. It does
not change the client connection, application-owned tools, MCP server configuration, or permission
boundary.
Return to Step 7: Run and explain the application.