Comfy-Org/ComfyUI · error · ValueError

'thinking' can only be disabled for text-to-image; enable it

Error message

'thinking' can only be disabled for text-to-image; enable it when using reference images.

What it means

Raised by ByteDanceSeedreamNodeV2 when the 'thinking' toggle is disabled (advanced input, default True) while reference images are connected. Seedream's prompt-optimization reasoning can only be turned off for pure text-to-image requests; image-conditioned requests must keep it enabled.

Source

Thrown at comfy_api_nodes/nodes_bytedance.py:1010

                )
            if out_num_pixels > 16_777_216:
                raise ValueError(
                    f"Maximum image resolution for the selected model is 16.78MP, but {mp_provided:.2f}MP provided."
                )

        image_tensors: list[Input.Image] = [t for t in images_dict.values() if t is not None]
        n_input_images = sum(get_number_of_images(t) for t in image_tensors)
        max_num_of_images = 14 if model_id == "seedream-5-0-260128" else 10
        if n_input_images > max_num_of_images:
            raise ValueError(
                f"Maximum of {max_num_of_images} reference images are supported, but {n_input_images} received."
            )
        if sequential_image_generation == "auto" and n_input_images + max_images > 15:
            raise ValueError(
                "The maximum number of generated images plus the number of reference images cannot exceed 15."
            )
        if not thinking and n_input_images > 0:
            raise ValueError(
                "'thinking' can only be disabled for text-to-image; enable it when using reference images."
            )

        reference_images_urls: list[str] = []
        if image_tensors:
            for tensor in image_tensors:
                validate_image_aspect_ratio(tensor, (1, 3), (3, 1))
            reference_images_urls = await upload_images_to_comfyapi(
                cls,
                image_tensors,
                max_images=n_input_images,
                mime_type="image/png",
                wait_label="Uploading reference images",
            )

        optimize_prompt_options = None
        if n_input_images == 0:
            optimize_prompt_options = Seedream5OptimizePromptOptions(thinking="enabled" if thinking else "disabled")

View on GitHub (pinned to 1c6d8d45b3)

Solutions

Re-enable 'thinking' (leave it True) whenever reference images are connected.
Remove the reference images if you truly want thinking disabled.
Keep thinking at its default True unless the request is pure text-to-image.

Example fix

// before
thinking: false
images: [reference.png] // -> raises

// after
thinking: true
images: [reference.png] // passes

Defensive patterns

Strategy: validation

Validate before calling

if image_tensors:  # any reference images connected
    thinking = True  # force-enable; cannot be off for i2i

Prevention

Treat thinking=True as the only valid setting in image-conditioned workflows.
Audit node configs copied between t2i and i2i graphs.

When it happens

Trigger: Connecting any image input (model dict 'images' with non-None tensors) and setting thinking=False in the node's advanced settings.

Common situations: Turning off thinking to save time on a text-to-image workflow, then adding a reference image later without re-enabling it; copying a node configuration between t2i and i2i workflows.

Related errors

AI-assisted analysis of Comfy-Org/ComfyUI@1c6d8d45b3 (2026-08-14). Data as JSON: /api/errors/96a375f70497e628. Report an issue: GitHub.