The U.S. General Services Administration (GSA) has launched a new initiative with Meta to expand access to Meta’s Llama open source AI models across federal departments and agencies. This effort is a key component of the GSA’s OneGov strategy, which streamlines government operations by eliminating redundant negotiations and enabling faster, unified adoption of technology across the federal landscape. Building on prior efforts that made Llama available to government agencies and contractors working on national security projects—including the historic deployment of Llama into space aboard the International Space Station National Laboratory—this new agreement marks a significant step forward in accelerating AI adoption throughout the federal government. It supports the goals of America’s AI Action Plan and aligns with key Office of Management and Budget (OMB) directives, including M-25-21, which calls for accelerating federal use of AI through innovation, governance, and public trust, and M-25-22, which focuses on efficient acquisition of AI technologies. The Llama models provide federal agencies with the ability to maintain full control over their data, including processing and storage, which is critical for compliance with security and privacy standards. As publicly available models, Llama enables technical teams to develop, deploy, and scale AI applications at a lower cost, delivering substantial value to American taxpayers. “Thanks to Meta and its commitment to advancing public sector services, our federal partners now have easier access to Llama and open source AI—tools that offer unique advantages for government use cases,” said Federal Acquisition Service Commissioner Josh Gruenbaum. “Through these OneGov initiatives, GSA is driving an unprecedented acceleration of AI adoption across the federal government.” Unlike traditional procurement agreements, this arrangement did not require individual agency negotiations because the Llama models are freely available. Instead, the GSA focused on verifying that Llama meets federal security, compliance, and operational requirements, and ensured consistent, secure access across agencies. This collaboration strengthens the federal government’s ability to experiment with, adapt, and deploy AI solutions while safeguarding sensitive data. It empowers agencies to build mission-specific applications, reduces reliance on closed-source AI providers, and promotes transparency and reproducibility in AI development. By embracing open source technology, the federal government is not only improving efficiency but also reinforcing its leadership in responsible AI innovation.