Amazon SageMaker JumpStart adds model discovery and one-click Hugging Face deployment

Amazon's AWS Machine Learning Blog shows how to deploy a pretrained model from SageMaker JumpStart, run timed sample prompts with the SageMaker Python SDK, and delete the endpoint afterward.

In a sample run, a model returned a 153-word answer in 9.07 seconds, about 16.9 words per second. The post says 2026 updates let users browse the catalog in the Model Hub and deploy Hugging Face models in one click.