US Government Vets OpenAI's GPT-5.6 Launch in Major AI Policy Shift

A New Frontier in AI Regulation

In a significant departure from its initial laissez-faire stance, the Trump administration has enacted a decisive new policy toward frontier artificial intelligence. The U.S. government will now directly vet which companies gain initial access to OpenAI's newly announced GPT-5.6 models, a move that marks the most substantial federal intervention into AI model releases to date.

This policy shift follows similar restrictions placed on Anthropic's advanced Fable 5 and Mythos 5 models earlier this month. According to sources, the White House's Office of the National Cyber Director and the Office of Science and Technology Policy formally requested OpenAI limit the GPT-5.6 rollout, citing unresolved national security concerns.

The development signals Washington is beginning to treat the most powerful U.S.-developed AI models as dual-use technologies requiring government review before public release. OpenAI CEO Sam Altman has been in direct discussions with Commerce Secretary Howard Lutnick and other administration officials to coordinate the controlled launch.

GPT-5.6: A Trio of Models Under Lock

OpenAI officially announced the GPT-5.6 series on Friday, introducing three distinct models. The flagship model is named Sol, designed for maximum capability. Alongside it are Terra, optimized for daily work tasks, and Luna, a more cost-effective version.

Despite the launch, access is severely restricted. OpenAI stated the models are available only in a "limited preview for a small group of trusted partners" approved by the U.S. government. Initial reports suggest the launch may be limited to roughly 20 government-approved partners.

The company emphasized this is a temporary measure, with plans for a broader release in the coming weeks pending government evaluation. OpenAI has been previewing GPT-5.6 capabilities with federal agencies for the past month, including in White House meetings in early June.

Industry Pushback and a Reluctant Compliance

OpenAI's cooperation comes with notable public reservations. In a memo to employees obtained by The Information, Altman stated, "We've made clear to the U.S. government that this is not our preferred long term model." The company echoed this in an official blog post, arguing such processes "keep the best tools from users, developers, enterprises, cyber defenders, and global partners who need them."

However, OpenAI framed the restrictions as a necessary short-term step. The company is operating in a regulatory gap: the administration has announced a plan to evaluate new AI releases but has not yet finalized the testing framework. OpenAI believes compliance now is the fastest path to wider availability.

"We are taking this short-term step because we believe it is the strongest path to broader availability in the coming weeks," the company stated, while it works with the administration to develop a "repeatable process for future model releases."

The Broader Market and Regulatory Context

This move places OpenAI in a similar position to its competitor, Anthropic, which saw its frontier models halted by a rare Commerce Department directive. The parallel actions suggest a coordinated, sector-wide approach is emerging from Washington.

Analysts note this transforms the AI race. It is no longer solely about technological superiority but also about navigating an evolving regulatory landscape. The requirement for separate government permission for each user introduces a new layer of complexity and potential delay to product launches.

While Altman estimates the vetting process will take only a few weeks, the experience with Anthropic's models suggests these periods could stretch into months. This creates economic risks and uncertainty for AI firms reliant on rapid iteration and deployment.

Implications for the Future of AI Development

The U.S. government's preemptive vetting of GPT-5.6 represents a foundational shift in AI governance. It establishes a precedent where advanced AI models are treated with a level of oversight previously reserved for exports with clear military applications.

For the industry, this introduces a new variable in strategic planning. The timeline between model development and commercial deployment now includes an unpredictable government review period. This could advantage well-established firms with existing government relations while creating barriers for smaller startups.

For the global AI landscape, it signals that the United States is willing to slow the domestic release of its most powerful technology to assess risks. This contrasts with the traditional Silicon Valley ethos of moving fast and iterating publicly, potentially affecting the pace of innovation.

The coming weeks will be critical. The administration's evaluation of GPT-5.6 and the subsequent framework it develops will set the template for how all future frontier AI models from U.S. companies are brought to market. The balance struck between security and openness will define the next era of the AI race.