model of create_session not working with system prompt

[zeerd]

#109
#115

I am sure of that I used gpt-4.1 as model for create_session.
But my premium request. runs out.

root@df432b7e7748:/workspace/code/.build-test# cat a.py
import asyncio
from copilot import CopilotClient

async def main():
    # Create and start client
    client = CopilotClient()
    await client.start()

    # Create a session
    session = await client.create_session({"model": "gpt-4.1"})

    # Wait for response using session.idle event
    done = asyncio.Event()

    def on_event(event):
        if event.type.value == "assistant.message":
            print(event.data.content)
        elif event.type.value == "session.idle":
            done.set()

    session.on(on_event)

    # Send a message and wait for completion
    await session.send({"prompt": "Who are you and what model you are?"})
    await done.wait()

    # Clean up
    await session.destroy()
    await client.stop()

asyncio.run(main())
root@df432b7e7748:/workspace/code/.build-test# python a.py
I am GitHub Copilot CLI, an AI-powered terminal assistant designed to help with software engineering tasks. My current model is Claude Sonnet 4.6.

GitHub Copilot CLI 0.0.421.
github-copilot-sdk==0.1.25

Edited:
With 0.1.30, sometimes the model correct. But has a chance to back to Claude Sonnet 4.6, too.

[zeerd]

with 0.1.30, this issue occur when replaced the system prompt.

import asyncio
from copilot import CopilotClient
from copilot.types import PermissionHandler


async def main():
    # Create and start client
    client = CopilotClient()
    await client.start()

    # Create a session
    session = await client.create_session({
        "model": "gpt-4.1",
        "on_permission_request": PermissionHandler.approve_all,
        "system_message": {
            "mode": "replace",
            "content": "You are a helpful assistant for C++ Unit Test."
        }
    })
    await session.set_model("gpt-4.1")

    # Wait for response using session.idle event
    done = asyncio.Event()

    def on_event(event):
        if event.type.value == "assistant.message":
            print(event.data.content)
        elif event.type.value == "session.idle":
            done.set()

    session.on(on_event)

    # Send a message and wait for completion
    await session.send({
        "prompt": (
            "Who are you and what model you are "
            "and what kind of assistant you are?"
        )
    })
    await done.wait()

    # Clean up
    await session.destroy()
    await client.stop()

asyncio.run(main())

I am an AI-powered assistant designed to help you with C++ unit testing and related development tasks. My main capabilities include:

- Explaining C++ code and unit testing concepts
- Generating and reviewing C++ unit tests (using frameworks like Google Test, Catch2, etc.)
- Debugging test failures and suggesting fixes
- Automating test setup and execution
- Assisting with code quality and best practices

Model: I am currently running on the Claude Sonnet 4.6 model, which is optimized for code understanding, reasoning, and technical assistance.

If you have any questions or need help with C++ unit tests, just let me know!