autogen - getting started with gemini API

By default code initialization using Open API client, would uses OPEN API platform  https://platform.openai.com/account/api-keys. 

To use Gemini, we have to initialize it using the following code to avoid issues. (otherwise it goes back to OPEN API endpoints instead of google AI endpoints (https://ai.google.dev)

from autogen_agentchat.agents import AssistantAgent

from autogen_agentchat.ui import Console

from autogen_ext.models.openai import OpenAIChatCompletionClient

import asyncio

from autogen_core.models import ModelInfo

from autogen_core.models import UserMessage

# Run the agent and stream the messages to the console.

async def main() -> None:

    # model_client = OpenAIChatCompletionClient(

    # model="gemini-1.5-flash-8b",

    # api_key="AIzaSyAZPjwmY7e5Ti9NKHzjPjbYQpx1dmPwLI8",

    # )

    model_client = OpenAIChatCompletionClient(

    model="gemini-2.0-flash-lite",

    model_info=ModelInfo(vision=True, function_calling=True, json_output=True, family="unknown", structured_output=True),  

    api_key="AIzaSyAZPjwmY7e5Ti9NKHzjPjbYQpx1dmPwLI8",

    )

    response = await model_client.create([UserMessage(content="What is the capital of France?", source="user")])

    print(response)

    await model_client.close()

if __name__ == "__main__":

    asyncio.run(main())

You can find more models here

# https://ai.google.dev/gemini-api/docs/models