The U.S. Department of Commerce has cleared OpenAI for a broad release of its advanced GPT-5.6 model, a move that consolidates control over a technology poised to reshape society. OpenAI expects the rollout as early as this week, following additional testing and meetings with government officials. This decision, reported by Axios and Reuters, signals a new era of government-corporate oversight on critical digital infrastructure.
OpenAI announced late Tuesday that its flagship GPT-5.6 model, Sol, alongside lower tiers Terra and Luna, will launch publicly on Thursday. This launch follows a delay last month, prompted by U.S. government requests citing heightened national security concerns over potential misuse of powerful AI technologies. CNBC confirmed OpenAI's anticipation of this week's rollout.
Elite Oversight and Secrecy
Testing of the models was conducted by the Center for AI Standards and Innovation, a division within the Department of Commerce. OpenAI dispatched technical experts who remained in Washington, D.C., to address any questions. Reuters noted that OpenAI initially restricted GPT-5.6 access to a small group of vetted partners, whose details were shared with authorities. CNBC added that OpenAI did not disclose the names of these partners, ensuring compliance with federal oversight remained opaque.
Last month, the Trump administration pushed OpenAI for a staggered release of GPT-5.6, limiting initial access to government-approved entities. OpenAI stated at the time that this staggered approach was not its preferred method for model releases. The company also acknowledged that AI firms and the government are operating without finalized concrete standards for such models, despite calls for them in President Trump's latest AI executive order.
The National Security Pretext
Washington has intensified its scrutiny of advanced AI model releases, citing the need to identify potential threats. Concerns center on the technology's possible misuse by military or intelligence establishments in China, Russia, and other nations. Reuters highlighted the ongoing race between the United States and China to develop cutting-edge AI, warning of its potential to dramatically accelerate sophisticated cyberattacks across complex, interconnected, and often outdated technology systems.
OpenAI previewed the models in late June, touting improved agentic capabilities in coding, biology, and cybersecurity. GPT-5.6 Sol, the company claimed, was competitive with Anthropic's Mythos Preview on the ExploitBench cybersecurity benchmark. Meanwhile, Elon Musk, whose SpaceXAI rivals Anthropic and OpenAI, announced on Wednesday that his company was making its leading model, Grok 4.5, available to the public.
Unaccountable Power
President Donald Trump previously signed an executive order establishing a voluntary framework for AI developers. This framework allows developers to provide "covered frontier models" to the U.S. government for up to 30 days before releasing them to trusted partners. CNBC reported that this government decision reflects the Trump administration's increasingly hands-on approach to AI regulation, aiming to assess model capabilities prior to full-scale release.
This regulatory muscle was demonstrated when OpenAI competitor Anthropic abruptly disabled its most advanced AI models, Mythos 5 and Fable 5, for all users. This action followed the U.S. government's June 12 export control order, again citing national security concerns. The curbs were lifted last week, Reuters confirmed, after Anthropic implemented certain safeguards, ending a period of regulatory uncertainty that limited global availability. In contrast, Zhipu, trading as Knowledge Atlas Technology JSC, launched its GLM 5.2 model last month, which remains free to download, fine-tune, and run on enterprise servers. The White House, Department of Commerce, and OpenAI did not respond to requests for comment regarding these developments.