[BUG] SageMakerAIModel does not support SageMaker Async Inference endpoints

[tomo-vzc]

Checks

• I have updated to the lastest minor and patch version of Strands
• I have checked the documentation and this is not expected behavior
• I have searched ./issues and there are no duplicates of my issue

Strands Version

1.33.0

Python Version

3.13.12

Operating System

Ubuntu

Installation Method

pip

Steps to Reproduce

1. Deploy any model to a SageMaker endpoint with AsyncInferenceConfig:

from sagemaker.core.inference_config import AsyncInferenceConfig

async_config = AsyncInferenceConfig(
   output_path="s3://my-bucket/async-outputs/",
)

model_builder.deploy(
   instance_type="ml.g5.xlarge",
   initial_instance_count=1,
   endpoint_name="my-async-endpoint",
   inference_config=async_config,
)

2. Use SageMakerAIModel to invoke it:

from strands import Agent
from strands.models.sagemaker import SageMakerAIModel

model = SageMakerAIModel(
    endpoint_config={
        "endpoint_name": "my-async-endpoint",
        "region_name": "us-west-2",
    },
    payload_config={"max_tokens": 100, "stream": False},
)

agent = Agent(model=model)
result = agent("Say hello.")  # raises ValidationError

3. Both stream=True and stream=False fail with ValidationError

Expected Behavior

Responce

Actual Behavior

FAILED: ValidationError: An error occurred (ValidationError) when calling the InvokeEndpoint operation: Endpoint XXXX does not support this inference type.

Motivation

SageMaker Async Inference seems to me the recommended approach for long-running inference (>60s), which is common for large LLMs with long prompts. The 60-second real-time timeout on invoke_endpoint makes it unusable for many production workloads.