Gemini as model client with autogen 0.5

[sonnylaskar]

What happened?

Describe the bug
I am referring to https://microsoft.github.io/autogen/stable/user-guide/agentchat-user-guide/tutorial/models.html#gemini-experimental but the documentation seems incomplete

To Reproduce

from autogen_ext.models.openai import OpenAIChatCompletionClient
model_client = OpenAIChatCompletionClient(
...     model="gemini-2.0-flash-lite",
...     api_key='<>',
... )

..._model_info.py", line 314, in get_info
    raise ValueError("model_info is required when model name is not a valid OpenAI model")
ValueError: model_info is required when model name is not a valid OpenAI model

How do I get it to work with Gemini.
I am using v0.5.1

Which packages was the bug in?

Python Extensions (autogen-ext)

AutoGen library version.

Python 0.5.1

Other library version.

No response

Model used

gemini-2.0-flash-lite

Model provider

Google Gemini

Other model provider

No response

Python version

3.12

.NET version

None

Operating system

None

[victordibia]

@sonnylaskar ,

Thanks for the issue.

The list of Gemini models supported explicitly is here

For any model not on that list (e.g, flash-lite which you are exploring), then you will need to provide some model information (e.g., if it has support for vision, structured output etc ... this information is used by the AutoGen framework).

from autogen_core.models import UserMessage
from autogen_ext.models.openai import OpenAIChatCompletionClient
from autogen_core.models import ModelInfo

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="GEMINI_API_KEY",
)

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

[sonnylaskar]

Thank you @victordibia for the super crazy fast response 👍
This is exactly what I was looking for.

[shijianzhong]

but，in autoGen studio,how to add the config

[victordibia]

@shijianzhong ,

It would look like this (the model client section of the team configuration).
See related docs.

{
      "provider": "autogen_ext.models.openai.OpenAIChatCompletionClient",
      "component_type": "model",
      "version": 1,
      "component_version": 1,
      "description": "Gemini 2.5",
      "label": "Gemini 2.5",
      "config": {
        "model": "gemini-2.0-flash-lite",
        "model_info": {
          "vision": false,
          "function_calling": true,
          "json_output": false,
          "family": "unknown",
          "structured_output": false
        },
        "base_url": "https://generativelanguage.googleapis.com/v1beta/openai/",
        "api_key": "..."
      }
    }