The general availability of G6e instances on SageMaker provides users with access to up to 8 NVIDIA L40 Tensor Core GPUs, each with 48 GB memory, paired with third-generation AMD EPYC CPUs. These instances offer up to 2.5 times improved performance over earlier G5 generations, facilitating demanding workloads such as generative AI fine-tuning, large language model deployment, and media generation.
Available in multiple AWS regions, these instances enable researchers and developers to experiment interactively with models containing up to 13 billion parameters. This upgrade aids in reducing training times and scaling AI model experimentation within the SageMaker notebook environment, supporting advanced computing use cases effectively.
Source & further reading
SageMaker Notebook Instances now support G6e instance types ↗Source published 2026-06-23. 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.