BerriAI/litellm · error · ValueError

Unknown BFL image edit model: {model_name}. Supported models

Error message

Unknown BFL image edit model: {model_name}. Supported models: {list(IMAGE_EDIT_MODELS.keys())}

What it means

The BFL image-edit transformation maps model names to endpoints via the IMAGE_EDIT_MODELS dict. After lowercasing and stripping any provider prefix (e.g. 'black_forest_labs/flux-kontext-pro' -> 'flux-kontext-pro'), an unknown name raises plain ValueError listing the supported models. It fires before any request is sent, so no API quota is consumed.

Source

Thrown at litellm/llms/black_forest_labs/image_edit/transformation.py:171

        """
        BFL uses JSON requests, not multipart/form-data.
        """
        return False

    def _get_model_endpoint(self, model: str) -> str:
        """
        Get the API endpoint for a given model.
        """
        # Remove provider prefix if present (e.g., "black_forest_labs/flux-kontext-pro")
        model_name = model.lower()
        if "/" in model_name:
            model_name = model_name.split("/")[-1]

        # Check if model is in our mapping
        if model_name in IMAGE_EDIT_MODELS:
            return IMAGE_EDIT_MODELS[model_name]

        raise ValueError(
            f"Unknown BFL image edit model: {model_name}. Supported models: {list(IMAGE_EDIT_MODELS.keys())}"
        )

    def get_complete_url(
        self,
        model: str,
        api_base: str | None,
        litellm_params: dict,
    ) -> str:
        """
        Get the complete URL for the Black Forest Labs API request.
        """
        base_url: str = api_base or get_secret_str("BFL_API_BASE") or DEFAULT_API_BASE
        base_url = base_url.rstrip("/")

        endpoint: Final = self._get_model_endpoint(model)
        return f"{base_url}{endpoint}"

View on GitHub (pinned to 6c2dcb801b)

Solutions

Use a model present in IMAGE_EDIT_MODELS — typically 'black_forest_labs/flux-kontext-pro' or 'black_forest_labs/flux-kontext-max' (the error message prints the exact supported list; match it).
If you intended text-to-image rather than editing, call litellm.image_generation with the generation model instead.
Upgrade LiteLLM if BFL shipped a new edit model that your version does not map yet.
Check for typos and stray whitespace in the model string.

Example fix

# before
litellm.image_edit(model="black_forest_labs/flux-pro-1.1", image=img, prompt="...")

# after
litellm.image_edit(model="black_forest_labs/flux-kontext-pro", image=img, prompt="...")

Defensive patterns

Strategy: validation

Validate before calling

from litellm.llms.black_forest_labs.image_edit.transformation import IMAGE_EDIT_MODELS

def norm(m): return m.lower().split("/")[-1]
assert norm(model) in IMAGE_EDIT_MODELS, f"{model} not in {list(IMAGE_EDIT_MODELS)}"

Type guard

def is_bfl_edit_model(model: str) -> bool:
    return model.lower().split("/")[-1] in IMAGE_EDIT_MODELS

Try / catch

try:
    litellm.image_edit(model=model, image=img, prompt=p)
except ValueError as e:
    if "Unknown BFL image edit model" in str(e):
        model = "black_forest_labs/flux-kontext-pro"  # fall back to known-good
        litellm.image_edit(model=model, image=img, prompt=p)
    else:
        raise

Prevention

Keep a module-level constant of supported edit models instead of raw strings.
Pin model names from IMAGE_EDIT_MODELS, not from marketing docs.
Test model strings in CI with a no-network unit test that calls _get_model_endpoint.

When it happens

Trigger: Calling image_edit with model set to a name not in IMAGE_EDIT_MODELS — e.g. a generation-only model like 'flux-pro-1.1' used with image_edit, a typo such as 'flux-kontext-proo', or an unlisted casing variant (matching is lowercase).

Common situations: Confusing image generation models with image edit models (only the kontext family supports edit); using a model string from BFL docs that LiteLLM has not mapped yet (version skew after BFL releases new models); copy-paste typos; omitting or mistyping the provider prefix is fine, but the suffix must match exactly.

Related errors

AI-assisted analysis of BerriAI/litellm@6c2dcb801b (2026-08-15). Data as JSON: /api/errors/96a2ad30ea0da451. Report an issue: GitHub.