Argonne launches high-performance computing-backed AI research service

Argonne National Laboratory introduces a new AI inference service designed to provide researchers with access to a variety of artificial intelligence models.
The service aims to support advanced research and innovation in AI, aligning with the Department of Energy’s mission.
The platform will offer large language models and scientific foundation models, reducing the need for researchers to build and maintain their own AI infrastructure.
Models available include open-weight models and domain-specific science foundation models.
The inference service is powered by Argonne’s flagship exascale computer, Aurora, and other hardware to facilitate rapid data analysis.
Researchers at various national laboratories, including Los Alamos, Brookhaven, Lawrence Berkeley, Fermi, Lawrence Livermore, Oak Ridge, and Sandia, can access and benefit from the service.
The platform can support applications beyond AI research, such as fusion energy, chemistry, and materials science.
It builds upon a framework developed in 2025 to enable the running of multiple AI tasks in parallel, with the ultimate aim of advancing the Genesis Mission.