Users can adapt these large foundation models using supervised fine-tuning, direct preference optimization, and reinforcement fine-tuning, with options for low-rank adaptation or full model tuning. This expands the portfolio of models available for serverless training in SageMaker AI alongside previously supported families such as Nova and Gemma.
Serverless customization removes infrastructure management burdens, allowing customers to focus on data and evaluation while paying only for usage. The feature is available in multiple AWS regions and supports integration via SageMaker Studio, Python SDK, and coding assistants, facilitating tailored model deployment for diverse AI applications.
Source & further reading
Amazon SageMaker AI now supports serverless model customization for Muse Glimmer, Qwen 3.8 and Ministral 3 models ↗Source published 2026-10-22. This report was prepared with AI assistance from the linked primary-source material. Claims and limitations should be checked against the original announcement or research.