AssistantAgent with OpenRouter can't function call if there is a output_content_type

[ChenFryd]

What happened?

Describe the bug
Open Router models can't produce function calls when have structred output.

To Reproduce

from autogen_agentchat.agents import AssistantAgent
from autogen_agentchat.ui import Console
from autogen_ext.models.openai import OpenAIChatCompletionClient
from autogen_core.models import ModelFamily

# Define a model client. You can use other model client that implements
# the `ChatCompletionClient` interface.
from autogen_core.tools import FunctionTool
from dotenv import load_dotenv
import os
load_dotenv()
from pydantic import BaseModel, Field

class weather(BaseModel):
    city: str = Field(
        ...,
        description="The city we get weather from"
    )
    temperature: int = Field(
        ...,
        description="Temperature in Farenheits"
    )

model_client = OpenAIChatCompletionClient(
        base_url=os.getenv("OPEN_ROUTER_BASE_URL"),
        api_key=os.getenv("OPEN_ROUTER_API_KEY"),
        model="openai/gpt-oss-20b:free",
        model_info={
            "vision": False,
            "function_calling": True,
            "json_output": True,
            "family": ModelFamily.UNKNOWN,
            "structured_output": True,
        }
    )


# Define a simple function tool that the agent can use.
# For this example, we use a fake weather tool for demonstration purposes.
async def get_weather(city: str) -> str:
    """Get the weather for a given city."""
    return f"The weather in {city} is 73 degrees and Sunny."


# Define an AssistantAgent with the model, tool, system message, and reflection enabled.
# The system message instructs the agent via natural language.

agent = AssistantAgent(
    name="weather_agent",
    model_client=model_client,
    tools=[FunctionTool(get_weather,description="get weather", strict=True)],
    output_content_type=weather,
    system_message="You are a helpful assistant. Use your function calls",
    reflect_on_tool_use=True,
    model_client_stream=True,  # Enable streaming tokens from the model client.
)


# Run the agent and stream the messages to the console.
async def main() -> None:
    await Console(agent.run_stream(task="What is the weather in New York?"))
    # Close the connection to the model client.
    await model_client.close()


# NOTE: if running this inside a Python script you'll need to use asyncio.run(main()).
import asyncio
asyncio.run(main())

Expected behavior
The agent to return
`{"city":"New York","temperature":73}

But it does not call the tool, so it will return whatever temperature

Screenshots
If applicable, add screenshots to help explain your problem.

Additional context
gpt-5-nano does able to use this function call

Which packages was the bug in?

Python AgentChat (autogen-agentchat>=0.4.0)

AutoGen library version.

Python dev (main branch)

Other library version.

No response

Model used

gpt-oss-20b, all open source models

Model provider

Other (please specify below)

Other model provider

All Open Source models

Python version

3.12

.NET version

None

Operating system

Ubuntu