BerriAI/litellm · error · BedrockError

Bedrock Invoke HTTPX: Unknown provider={provider}, model={mo

Error message

Bedrock Invoke HTTPX: Unknown provider={provider}, model={model}. Try calling via converse route - `bedrock/converse/<model>`.

What it means

Raised by BaseAmazonInvokeConfig.transform_request when the provider segment parsed from the model string (bedrock/<provider>.<model>) does not match any known Invoke-API transformation (anthropic, cohere, ai21, meta/llama, mistral, amazon, openai, deepseek_r1, twelvelabs). It is a BedrockError 404 with a message that suggests using the Converse route instead.

Source

Thrown at litellm/llms/bedrock/chat/invoke_transformations/base_invoke_transformation.py:269

        elif provider == "twelvelabs":
            return litellm.AmazonTwelveLabsPegasusConfig().transform_request(
                model=model,
                messages=messages,
                optional_params=optional_params,
                litellm_params=litellm_params,
                headers=headers,
            )
        elif provider == "openai":
            # OpenAI imported models use OpenAI Chat Completions format
            return litellm.AmazonBedrockOpenAIConfig().transform_request(
                model=model,
                messages=messages,
                optional_params=optional_params,
                litellm_params=litellm_params,
                headers=headers,
            )
        else:
            raise BedrockError(
                status_code=404,
                message=f"Bedrock Invoke HTTPX: Unknown provider={provider}, model={model}. Try calling via converse route - `bedrock/converse/<model>`.",
            )

        return request_data

    def transform_response(
        self,
        model: str,
        raw_response: httpx.Response,
        model_response: ModelResponse,
        logging_obj: LiteLLMLoggingObj,
        request_data: dict,
        messages: list[AllMessageValues],
        optional_params: dict,
        litellm_params: dict,
        encoding: Any,
        api_key: str | None = None,

View on GitHub (pinned to 6c2dcb801b)

Solutions

1. Switch the model string to bedrock/converse/<model> as the message suggests - Converse handles all providers uniformly.
2. Upgrade litellm to a version that maps the new provider.
3. Fix typos in the provider segment of the model ID (e.g. bedrock/meta.llama3-..., bedrock/cohere.command-...).
4. For imported OpenAI models on Bedrock use the bedrock/openai.* naming supported by recent litellm.

Example fix

# before
resp = litellm.completion(model="bedrock/<newprovider>.<new-model>", messages=msgs)

# after
resp = litellm.completion(model="bedrock/converse/<newprovider>.<new-model>", messages=msgs)

KNOWN_PROVIDERS = {"anthropic", "cohere", "ai21", "meta", "llama", "mistral", "amazon", "openai", "deepseek_r1", "twelvelabs"}
def normalize_bedrock_model(model: str) -> str:
    provider = model.split("/", 1)[-1].split(".", 1)[0]
    if provider not in KNOWN_PROVIDERS:
        return f"bedrock/converse/{model.split('/', 1)[1]}"  # unknown to invoke -> converse
    return model

def is_invoke_supported(model: str) -> bool:
    try:
        provider = model.split("/", 1)[1].split(".", 1)[0]
    except IndexError:
        return False
    return provider in {"anthropic", "cohere", "ai21", "meta", "llama", "mistral", "amazon", "openai", "deepseek_r1", "twelvelabs"}

from litellm.exceptions import BedrockError
try:
    resp = litellm.completion(model=model, messages=msgs)
except BedrockError as e:
    if e.status_code == 404 and "Unknown provider" in str(e):
        resp = litellm.completion(model=model.replace("bedrock/", "bedrock/converse/", 1), messages=msgs)
    else:
        raise

When it happens

Trigger: Using a model string like bedrock/myvendor.custom-model-v1 or bedrock/<unrecognized>.<model> through the legacy invoke path, including new providers (e.g. newly launched Bedrock models) not yet mapped in the installed litellm version.

Common situations: AWS launches a new provider/model before litellm adds the invoke transformer, typo in the model string provider segment, or attempting bedrock/openai.* without the openai mapping in older versions.

Related errors

• Error processing={raw_response.text}, Received error={e}
• Unexpected mistral completion response
• Error parsing received text={outputText}.\nError-{e}
• Model needs to be set for bedrock
• BedrockException: Context Window Error - {error_str}

AI-assisted analysis of BerriAI/litellm@6c2dcb801b (2026-08-15). Data as JSON: /api/errors/556747d3e412725a. Report an issue: GitHub.