microsoft/autogen · error · ValueError

Missing required field '{field}' in ModelInfo. Starting in v

Error message

Missing required field '{field}' in ModelInfo. Starting in v0.4.7, the required fields are enforced.

What it means

validate_model_info enforces that every ModelInfo dict passed to a ChatCompletionClient contains the keys vision, function_calling, json_output, and family — a contract made mandatory in autogen v0.4.7. Missing any one raises ValueError naming the field. A fifth key, structured_output, currently only emits a UserWarning but will become required in a future release.

Source

Thrown at python/packages/autogen-core/src/autogen_core/models/_model_client.py:194

    """True if the model supports json output, otherwise False. Note: this is different to structured json."""
    family: Required[ModelFamily.ANY | str]
    """Model family should be one of the constants from :py:class:`ModelFamily` or a string representing an unknown model family."""
    structured_output: Required[bool]
    """True if the model supports structured output, otherwise False. This is different to json_output."""
    multiple_system_messages: Optional[bool]
    """True if the model supports multiple, non-consecutive system messages, otherwise False."""


def validate_model_info(model_info: ModelInfo) -> None:
    """Validates the model info dictionary.

    Raises:
        ValueError: If the model info dictionary is missing required fields.
    """
    required_fields = ["vision", "function_calling", "json_output", "family"]
    for field in required_fields:
        if field not in model_info:
            raise ValueError(
                f"Missing required field '{field}' in ModelInfo. "
                "Starting in v0.4.7, the required fields are enforced."
            )
    new_required_fields = ["structured_output"]
    for field in new_required_fields:
        if field not in model_info:
            warnings.warn(
                f"Missing required field '{field}' in ModelInfo. "
                "This field will be required in a future version of AutoGen.",
                UserWarning,
                stacklevel=2,
            )


class ChatCompletionClient(ComponentBase[BaseModel], ABC):
    # Caching has to be handled internally as they can depend on the create args that were stored in the constructor
    @abstractmethod
    async def create(

View on GitHub (pinned to 027ecf0a37)

Solutions

1. Add all four required keys with correct booleans plus family, e.g. {"vision": False, "function_calling": True, "json_output": True, "family": "unknown", "structured_output": False}.
2. Copy a complete ModelInfo from a known model in autogen's built-in model lists and adjust the booleans for your endpoint.
3. Set family to "unknown" when the model is not a known family (ModelFamily.UNKNOWN).
4. Add structured_output now to silence the future-deprecation warning.

Example fix

# before
client = MyChatCompletionClient(
    model="my-model",
    model_info={"vision": False, "function_calling": True},  # ValueError: missing 'json_output'
)

# after
client = MyChatCompletionClient(
    model="my-model",
    model_info={
        "vision": False,
        "function_calling": True,
        "json_output": True,
        "structured_output": False,
        "family": "unknown",
    },
)

REQUIRED = ["vision", "function_calling", "json_output", "family"]

def complete_model_info(info: dict) -> dict:
    missing = [f for f in REQUIRED + ["structured_output"] if f not in info]
    if missing:
        raise ValueError(f"model_info missing fields: {missing}")
    return info

model_info = complete_model_info(model_info)

from typing import TypedDict

class CompleteModelInfo(TypedDict, total=True):
    vision: bool
    function_calling: bool
    json_output: bool
    structured_output: bool
    family: str

def is_complete_model_info(d: dict) -> bool:
    return all(k in d for k in CompleteModelInfo.__annotations__)

When it happens

Trigger: Creating a client with `model_info={"vision": False, "function_calling": False}` (family/json_output missing); carrying ModelInfo dicts written for autogen < 0.4.7 into a newer version; copying partial dicts from tutorials that predate the enforcement.

Common situations: Upgrading autogen-core/autogen-ext past v0.4.7 with existing client configs; custom or local model clients (e.g. OpenAI-compatible endpoints) where model_info was hand-written; LLM-generated config code that omits fields.

Related errors

• buffer_size must be greater than 0.
• head_size must be greater than 0.
• tail_size must be greater than 0.
• token_limit must be greater than 0.
• {cls.__name__} is a namespace class and cannot be instantiat

AI-assisted analysis of microsoft/autogen@027ecf0a37 (2026-08-15). Data as JSON: /api/errors/da2f6580472c7bd5. Report an issue: GitHub.