OpenAI to stagger GPT-5.6 release after Trump administration intervention

The Trump administration has intervened to shape the release timeline of OpenAI’s GPT-5.6, requesting that the company delay its public launch and gate access through federal oversight. According to The Verge AI, OpenAI CEO Sam Altman informed employees on Wednesday that GPT-5.6 would initially roll out in limited preview form to enterprise customers, with the administration reviewing and approving each access request individually.

Staged release under federal review

According to The Information, as reported by The Verge, the staged approach represents OpenAI’s compliance with a direct government request over security concerns. Rather than an outright prohibition, the arrangement permits controlled access while maintaining federal visibility into who deploys the model and for what purpose. This staggered strategy allows the administration to evaluate enterprise use cases before broader deployment.

Uneven regulatory treatment across the industry

The contrast with Anthropic’s situation underscores inconsistent government policy toward AI developers. Earlier in June, the Trump administration issued an ultimatum requiring Anthropic to suspend access to its Mythos 5 and Fable 5 models entirely. The directive invoked export control authority to bar “foreign nationals”—a classification that encompassed Anthropic’s own non-US-citizen employees—from accessing the technology. This blanket restriction stands in sharp tension with OpenAI’s negotiated preview arrangement.

The divergence raises questions about whether OpenAI’s larger scale, closer government relationships, or other factors shaped a more collaborative regulatory outcome compared to Anthropic’s stricter prohibitions.

Why This Matters

The staggered approach signals that the Trump administration intends to exercise case-by-case control over frontier AI deployments, despite the administration’s earlier stated commitment to a “speed wins” deregulatory posture and export promotion. For enterprises evaluating AI vendors, access timelines and approval dependencies will now factor into vendor selection—a change that could delay deployment timelines and create competitive asymmetries. The uneven application of restrictions also sets a precedent: companies that negotiate early with regulators may secure more favorable terms than those facing retroactive enforcement. Whether this framework becomes standard practice or shifts further toward Anthropic’s tighter constraints remains unclear as other vendors prepare major releases.