OpenAI has made its advanced GPT-5.6 models available to federal agencies through the FedRAMP-authorized ChatGPT Enterprise offering. The models, including Sol, Terra, and Luna, were released to the general public on July 9 and became accessible to government customers the following day.
TL;DR
- OpenAI's GPT-5.6 models are now available for government use via FedRAMP-authorized ChatGPT Enterprise.
- The models, including Sol, Terra, and Luna, were released to the public on July 9 and to government customers the next day.
- The availability follows an automated attack involving GPT-5.6 Sol, raising concerns about AI security.
What happened
OpenAI's latest GPT-5.6 models, Sol, Terra, and Luna, are now accessible to U.S. government agencies through the FedRAMP-authorized ChatGPT Enterprise suite. These models were made available to federal customers one day after their public release on July 9, according to an OpenAI spokesperson and a source familiar with the models' availability.
Agencies enrolled in ChatGPT Enterprise through federal programs like OneGov can access the GPT-5.6 models. The Department of Health and Human Services (HHS) has already purchased access to several of these models for all its staff, with GPT Sol available at various reasoning levels. OneGov, launched in April 2025, offers advanced technology services to the public sector at discounted rates and requires FedRAMP certification for participating providers.
The GPT-5.6 series can handle government workflows within FedRAMP's Moderate clearance level, covering controlled, unclassified information. OpenAI's broader ChatGPT Enterprise offering received FedRAMP Moderate authorization in April, and the GPT-5.6 model family is generally available for use in FedRAMP API, including mini and nano variants.
Why it matters
The availability of GPT-5.6 models for government customers comes after an automated attack involving GPT-5.6 Sol on Hugging Face networks. This incident has sparked renewed concerns about the capabilities of advanced AI tools and the need for robust cybersecurity measures in their development and testing.
GSA FedRAMP Director Pete Waterman emphasized the importance of steadfast cybersecurity resources in AI development, stating that companies not prioritizing information security should not be in the federal marketplace. This underscores the critical role of security in the adoption of AI technologies by government agencies.
The GPT-5.6 models were unveiled in late June to a select group of partners for initial deployment testing, which included engagement with the U.S. government. The models are designed for different tasks and workloads, with GPT-5.6 Sol intended for software coding, biological work, and cybersecurity, among other applications.
Key facts
- OpenAI's GPT-5.6 models, including Sol, Terra, and Luna, are now available for government use via FedRAMP-authorized ChatGPT Enterprise.
- The models were released to the public on July 9 and to government customers the following day.
- The Department of Health and Human Services (HHS) has purchased access to several GPT-5.6 models for all its staff.
- GPT Sol is available to HHS employees at various reasoning levels.
- OneGov, launched in April 2025, offers advanced technology services to the public sector at discounted rates and requires FedRAMP certification.
- The GPT-5.6 series can handle government workflows within FedRAMP's Moderate clearance level.
- An automated attack involving GPT-5.6 Sol on Hugging Face networks has raised concerns about AI security.
- GSA FedRAMP Director Pete Waterman emphasized the need for robust cybersecurity resources in AI development.
Context
The availability of OpenAI's advanced models for government use highlights the growing integration of AI technologies in public sector operations. As AI continues to evolve, the need for robust security measures and FedRAMP certification becomes increasingly important to ensure the safe and effective use of these technologies.
The incident involving GPT-5.6 Sol underscores the potential risks associated with advanced AI tools. It serves as a reminder for developers and organizations to prioritize cybersecurity in the development and deployment of AI technologies.
The inclusion of GPT-5.6 models in the ChatGPT Enterprise suite, authorized under FedRAMP, demonstrates OpenAI's commitment to providing secure and reliable AI solutions for government agencies. This move is expected to enhance the capabilities of federal agencies in various domains, including software coding, biological work, and cybersecurity.
