The Trump administration has requested OpenAI to limit GPT-5.6's launch to selected government partners for further evaluation.
Summary
- The Trump administration has requested that OpenAI limit the initial rollout of GPT-5.6 to select government-approved partners.
- Officials are currently evaluating the model under a new federal framework for advanced AI testing.
- This move follows ongoing demands from AI developers for greater oversight of sophisticated models.
The Trump administration has reportedly urged OpenAI to restrict the initial launch of its latest model, GPT-5.6, to a select group of government-sanctioned partners as federal officials conduct assessments. This request was first reported by multiple sources, including The Information and Axios.
This is the second time this month that the U.S. government has intervened to limit the release of an advanced AI model, following a directive to Anthropic to suspend public access to its models, Claude Fable 5 and Mythos 5, due to national security concerns.
According to reports, the White House's Office of the National Cyber Director and the Office of Science and Technology Policy have recommended that OpenAI confine the rollout of GPT-5.6 while the administration develops a framework to evaluate advanced AI models before they are made widely available. Sources suggest the request stems from the model's "Mythos-like" capabilities rather than signaling a broader shift in AI policy.
This latest action aligns with President Trump's recent executive order, which directed federal agencies to create a voluntary testing framework for advanced AI systems prior to their release, following extensive discussions about the program's structure.
The intervention highlights a shifting relationship between leading AI developers and the federal government, particularly as the development community has long advocated for regulatory measures in the industry. In recent Senate hearings, OpenAI CEO Sam Altman called for the establishment of an independent regulatory body to oversee advanced AI systems, emphasizing the need for oversight. Similarly, Anthropic CEO Dario Amodei argued that the most capable AI systems should undergo rigorous government-backed assessments before deployment, given their potential to contribute to advanced cyberattacks and other security threats.
The discussions around governance have gained momentum, with Anthropic, OpenAI, and Google proposing various frameworks for managing frontier AI. While their approaches differ, all three companies support structured evaluations of high-capability models, enhanced transparency in safety assessments, independent reviews of high-risk systems, and a more significant role for the government in AI oversight.
The administration's intervention may serve as a litmus test for the feasibility of the governance frameworks proposed by these leading AI firms, with critics warning that if regulations are influenced unevenly by dominant companies, it could result in regulatory capture, ultimately benefiting a select few and stifling competition.
