FoundationAgents/OpenManus · error · ValueError

Model {self.model} does not support images. Use a model from

Error message

Model {self.model} does not support images. Use a model from {MULTIMODAL_MODELS}

What it means

Raised by LLM.ask_with_images() when self.model is not in the MULTIMODAL_MODELS allowlist. The method attaches image_url content parts, which text-only models reject or silently mishandle, so it refuses up front. The allowlist is a module-level constant in this package, not provider capability discovery.

Source

Thrown at app/llm.py:519

            images: List of image URLs or image data dictionaries
            system_msgs: Optional system messages to prepend
            stream (bool): Whether to stream the response
            temperature (float): Sampling temperature for the response

        Returns:
            str: The generated response

        Raises:
            TokenLimitExceeded: If token limits are exceeded
            ValueError: If messages are invalid or response is empty
            OpenAIError: If API call fails after retries
            Exception: For unexpected errors
        """
        try:
            # For ask_with_images, we always set supports_images to True because
            # this method should only be called with models that support images
            if self.model not in MULTIMODAL_MODELS:
                raise ValueError(
                    f"Model {self.model} does not support images. Use a model from {MULTIMODAL_MODELS}"
                )

            # Format messages with image support
            formatted_messages = self.format_messages(messages, supports_images=True)

            # Ensure the last message is from the user to attach images
            if not formatted_messages or formatted_messages[-1]["role"] != "user":
                raise ValueError(
                    "The last message must be from the user to attach images"
                )

            # Process the last user message to include images
            last_message = formatted_messages[-1]

            # Convert content to multimodal format if needed
            content = last_message["content"]
            multimodal_content = (

View on GitHub (pinned to 52a13f2a57)

Solutions

1. Use a model from MULTIMODAL_MODELS, e.g. set model="gpt-4o" in config before calling ask_with_images()
2. If your model truly supports vision but isn't listed, add it to MULTIMODAL_MODELS in the module (or upgrade the package)
3. Branch your code: call ask() for text-only models and ask_with_images() only for vision models

Example fix

# before
llm = LLM(model="gpt-3.5-turbo")
await llm.ask_with_images("describe", [img])  # ValueError

# after
llm = LLM(model="gpt-4o")
await llm.ask_with_images("describe", [img])

from app.llm import MULTIMODAL_MODELS

def model_supports_images(model: str) -> bool:
    return model in MULTIMODAL_MODELS

from typing import TypeGuard
from app.llm import MULTIMODAL_MODELS

def is_vision_model(model: object) -> TypeGuard[str]:
    return isinstance(model, str) and model in MULTIMODAL_MODELS

try:
    out = await llm.ask_with_images("describe", images)
except ValueError as e:
    if "does not support images" in str(e):
        out = await llm.ask("describe (images omitted)")  # degrade to text
    else:
        raise

When it happens

Trigger: Constructing LLM with a text-only model (e.g. "gpt-3.5-turbo" or a non-vision fine-tune) then calling ask_with_images(); model name casing/alias mismatch causing the string to miss the allowlist; new vision model not yet added to MULTIMODAL_MODELS in this version.

Common situations: Switching the [llm] config to a cheaper text model and forgetting a code path calls ask_with_images; using a custom/local model name that is vision-capable but absent from the constant; version lag where a newly released vision model isn't in the list.

Related errors

• Unsupported image format: {image}
• The last message must be from the user to attach images
• No response received from the LLM
• Request may exceed input token limit (Current: {self.total_i
• Empty or invalid response from LLM

AI-assisted analysis of FoundationAgents/OpenManus@52a13f2a57 (2026-08-15). Data as JSON: /api/errors/4600a02b423e4206. Report an issue: GitHub.