Sitelet https://github.com/microsoft/autogen/pull/8077/files
Skip to content

fix(autogen-ext): add extra_body to OpenAI client config and serialization models #8077

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Muhammad Maaz (Maazkorejo) wants to merge 2 commits into
base: main
Choose a base branch
from
+64 −1
Open

fix(autogen-ext): add extra_body to OpenAI client config and serialization models #8077

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
1547361
fix(autogen-ext): add extra_body to OpenAI client config and serializ…
Maazkorejo Aug 19, 2026
ecdf0d4
test(autogen-ext): assert public dump_component payload and roundtrip…
Maazkorejo Aug 20, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 4 additions & 1 deletion python/packages/autogen-ext/src/autogen_ext/models/openai/config/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
from typing import Awaitable, Callable, Dict, List, Literal, Optional, Union
from typing import Any, Awaitable, Callable, Dict, List, Literal, Optional, Union

from autogen_core import ComponentModel
from autogen_core.models import ModelCapabilities, ModelInfo # type: ignore
Expand Down Expand Up @@ -57,6 +57,8 @@ class CreateArguments(TypedDict, total=False):
- 'low': Faster responses with less reasoning
- 'medium': Balanced reasoning and speed
- 'high': More thorough reasoning, may take longer"""
extra_body: Optional[Dict[str, Any]]
"""Additional vendor-specific or custom parameters to pass directly in the request body."""


AsyncAzureADTokenProvider = Callable[[], Union[str, Awaitable[str]]]
Expand Down Expand Up @@ -108,6 +110,7 @@ class CreateArgumentsConfigModel(BaseModel):
parallel_tool_calls: bool | None = None
# Controls the amount of effort the model uses for reasoning (reasoning models only)
reasoning_effort: Literal["minimal", "low", "medium", "high"] | None = None
extra_body: Dict[str, Any] | None = None


class BaseOpenAIClientConfigurationConfigModel(CreateArgumentsConfigModel):
Expand Down
60 changes: 60 additions & 0 deletions python/packages/autogen-ext/tests/models/test_openai_model_client.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3378,3 +3378,63 @@ async def test_reasoning_effort_validation() -> None:
}

ChatCompletionClient.load_component(config)


@pytest.mark.asyncio
async def test_extra_body_configuration_and_serialization() -> None:
"""Test extra_body parameter configuration, serialization, and deserialization."""
from autogen_core.models import ChatCompletionClient

# Test direct client initialization with extra_body
client = OpenAIChatCompletionClient(
model="gpt-4o",
api_key="fake_key",
extra_body={"enable_thinking": True, "custom_param": 123},
)
assert client._create_args["extra_body"] == {"enable_thinking": True, "custom_param": 123} # pyright: ignore[reportPrivateUsage]
assert client._raw_config["extra_body"] == {"enable_thinking": True, "custom_param": 123} # pyright: ignore[reportPrivateUsage]

# Test load_component with extra_body
config = {
"provider": "OpenAIChatCompletionClient",
"config": {
"model": "gpt-4o",
"api_key": "fake_key",
"extra_body": {"enable_thinking": True, "vendor_flag": "fast"},
},
}
loaded_client = ChatCompletionClient.load_component(config)
assert loaded_client._create_args["extra_body"] == {"enable_thinking": True, "vendor_flag": "fast"} # type: ignore[attr-defined] # pyright: ignore[reportPrivateUsage]
assert loaded_client._raw_config["extra_body"] == {"enable_thinking": True, "vendor_flag": "fast"} # type: ignore[attr-defined] # pyright: ignore[reportPrivateUsage]

# Test component dump payload inspection before reloading
dumped = client.dump_component()
assert dumped.config.extra_body == {"enable_thinking": True, "custom_param": 123}

# Test component dump and reload roundtrip
reloaded = OpenAIChatCompletionClient.load_component(dumped)
assert reloaded._create_args["extra_body"] == {"enable_thinking": True, "custom_param": 123} # pyright: ignore[reportPrivateUsage]
assert reloaded._raw_config["extra_body"] == {"enable_thinking": True, "custom_param": 123} # pyright: ignore[reportPrivateUsage]

# Verify the reloaded client's public dump matches original dump payload
reloaded_dumped = reloaded.dump_component()
assert reloaded_dumped.config.extra_body == {"enable_thinking": True, "custom_param": 123}

# Test AzureOpenAIChatCompletionClient dump and load roundtrip with extra_body
azure_client = AzureOpenAIChatCompletionClient(
model="gpt-4o",
azure_endpoint="https://fake.openai.azure.com/",
azure_deployment="gpt-4o-dep",
api_version="2024-06-01",
api_key="fake_key",
extra_body={"enable_thinking": True, "custom_param": 456},
)
assert azure_client._create_args["extra_body"] == {"enable_thinking": True, "custom_param": 456} # pyright: ignore[reportPrivateUsage]
azure_dumped = azure_client.dump_component()
assert azure_dumped.config.extra_body == {"enable_thinking": True, "custom_param": 456}

reloaded_azure = AzureOpenAIChatCompletionClient.load_component(azure_dumped)
assert reloaded_azure._create_args["extra_body"] == {"enable_thinking": True, "custom_param": 456} # pyright: ignore[reportPrivateUsage]
assert reloaded_azure.dump_component().config.extra_body == {"enable_thinking": True, "custom_param": 456}