selected_model returns None when session is created

[julianchun]

I am creating a session without explicitly passing a model, expecting the SDK to utilize a default model.
However, when I attempt to retrieve the model information from the session messages, the selected_model field is None.

Is this intended behavior?
If so, what is the recommended way to determine which default model the session is currently using before send a message?

import asyncio
from copilot import CopilotClient

async def main():
    client = CopilotClient()
    await client.start()
    session = await client.create_session() # without specify model
    messages = await session.get_messages()
    start_event = next(m for m in messages if m.type.value == "session.start")  
    print(f"selected model: {start_event.data.selected_model}") # selected model: None
  
if __name__ == "__main__":
    asyncio.run(main())

[github-actions]

Thanks for opening this issue @julianchun!

I've labeled this as a question about the sdk/python implementation.

This is a great question about default model behavior. The selected_model field being None in the session.start event when no explicit model is specified could be expected behavior if the model selection happens later (e.g., at message send time), or it could indicate that default model information isn't being propagated back to the session start event.

A maintainer will review this and clarify:

1. Whether this is the intended behavior
2. The recommended approach to determine which model will be used before sending messages

If you have any additional context (SDK version, CLI version, or other observations), feel free to add them!

AI generated by Issue Triage Agent