OpenAI has reportedly delayed the full public release of its next-generation AI model, GPT-5.6, following a formal request from the Trump administration to limit early access to a select group of government-approved entities.
This raises new concerns about the intersection of advanced AI capabilities and national security. According to a report by The Information on June 25, 2026, this decision reflects increasing federal involvement in the deployment of advanced AI systems, particularly those with significant cybersecurity implications.
The action comes amid heightened scrutiny of powerful AI models capable of automating complex tasks such as vulnerability discovery, exploit development, and planning multi-stage cyberattacks, indicating a shift toward tighter oversight before release.
According to CSN, the request follows a recent confrontation between the White House and the AI firm Anthropic. In this instance, the administration enforced export controls that temporarily halted the operations of its Fable 5 and Mythos 5 models.
These systems, especially Mythos 5, had been distributed under a controlled initiative called Project Glasswing to around 40 organizations, including major technology firms and financial institutions.
U.S. officials expressed concerns over Mythos’ capability to autonomously identify software vulnerabilities and execute chained attack sequences, abilities that could be weaponized if accessed by adversaries. This incident set a precedent for direct government intervention in the availability of AI models, particularly regarding dual-use cybersecurity functions.
Sources familiar with the situation indicate that both OpenAI and federal authorities view GPT-5.6 as comparable to Mythos in terms of advanced cyber capabilities.
This assessment led the Office of the National Cyber Director and the Office of Science and Technology Policy to formally request a staggered rollout, with access approvals being managed on a per-customer basis. Commerce Secretary Howard Lutnick reportedly advised against a broad release without coordination among agencies.
During an internal briefing, OpenAI CEO Sam Altman confirmed that the model would initially be available only to select enterprise partners as part of this controlled preview phase, with government agencies involved in vetting the recipients.
Although OpenAI has agreed to this arrangement, the company has expressed concerns about the long-term viability of such restrictions. Altman emphasized internally that, while the company is cooperating with the current government’s requests, it does not view this controlled-access model as sustainable for future releases.
The broader public launch of GPT-5.6 is expected within weeks, depending on the outcome of the preview phase and the government’s approval process. A White House official stated that federal agencies are continuing to collaborate with AI developers to establish shared safety and security practices as model capabilities rapidly evolve.
This development highlights a significant regulatory gap in the United States, where there is currently no formal framework for mandatory pre-release evaluations of advanced AI systems.
While President Trump’s executive order on AI innovation encourages voluntary sharing of models for cybersecurity review, compliance remains discretionary. The current arrangement with OpenAI, described as cooperative rather than legally binding, could represent an emerging template for public-private coordination.
The method of deploying GPT-5.6 may influence future policy decisions and shape how governments manage the risks associated with increasingly autonomous and security-sensitive AI technologies.
Follow us on Google News, LinkedIn, and X to Get Instant Updates and Set GBH as a Preferred Source in Google.
