What’s new here

This collection covers Moonshot AI’s Kimi models on AWS: one notebook for Bedrock, and links out to the AWS Marketplace listing and to SageMaker examples in another repo.

What it does

Three deployment paths are covered, but only the Bedrock path has a notebook in this repo:

• Amazon Bedrock: an example notebook walks through Kimi K3 via the Bedrock API. Older Kimi models are also available; the README links to AWS model cards for the current list.
• AWS Marketplace: points to the Kimi API Platform listing for direct access to Moonshot AI-operated endpoints.
• Self-hosted on AWS: links to external notebooks and a YAML config for deploying Kimi K3 on SageMaker HyperPod with EKS, and Kimi K2.5 on SageMaker AI with LMI.

Who it’s for

AWS builders who want to call Kimi models in their own accounts, whether through managed Bedrock APIs, a Marketplace subscription, or a self-hosted SageMaker cluster.

Try it

The README does not include a Quick Start command. To get started with Amazon Bedrock, open the example notebook linked in the README under “Kimi Models on Amazon Bedrock”. You need an AWS account with Amazon Bedrock access and a Python environment with Jupyter installed.

How mature is it

Created August 2026, last pushed October 2026. 7 stars, 2 forks, 3 contributors, 5 commits in the past 90 days. No releases. 3 open issues and pull requests. Licensed MIT-0.