OpenAI to Restrict GPT-5.6 Release at U.S. Government’s Request The debut of OpenAI’s most powerful GPT-5.6 AI models has turned into a live test of how far the U.S. government can go in controlling cutting‑edge algorithms before they reach the public.

In early June, President Trump signed a narrower AI executive order that asked leading labs to “voluntarily” give Washington up to 30 days of pre‑release access to particularly advanced systems, especially those with cyber capabilities. The order stopped short of formal licensing but signaled a shift from the administration’s earlier “hands off” posture.

As agencies built a security testing framework, the White House’s Office of the National Cyber Director and the Office of Science and Technology Policy quietly asked OpenAI to limit the initial rollout of GPT‑5.6 to a small set of government‑approved partners, marking the first time Washington had preemptively sought to restrict a model’s launch. Treasury, Commerce and other offices backed a staggered release to vet users.

On June 25, reports emerged that OpenAI CEO Sam Altman had told staff the administration would be “approving access customer by customer” during a preview period, with a broader release hoped for weeks later. That request followed harsher export‑control moves against rival Anthropic, which was ordered to cut off foreign‑national access to its Mythos and Fable frontier models.

Less than a day later, OpenAI formally unveiled GPT‑5.6 — a three‑model family of Sol, Terra, and Luna — but only in “limited preview” to roughly 20 enterprise partners whose participation had been cleared by the government. Sol, the flagship, is described as especially strong in coding, cybersecurity and biology, and is paired with new “max” and forthcoming “ultra” reasoning modes. OpenAI’s own system card emphasizes that GPT‑5.6 is backed by its “most robust” safeguards yet, designed to deliver the models “safely and at scale, around the world.”

OpenAI says the tight preview is being done at the government’s request under an existing Defense Department agreement and that it expects wider public access “in the coming weeks.” In a blog post and in Altman’s comments, the company has stressed that it does not want this to become standard. The current access process, it argued, “keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them,” even as the firm frames short‑term cooperation as the “strongest path” to broader availability while a repeatable federal review process is worked out.

From Washington’s side, officials present the intervention as a calibrated response to “Mythos‑like” capabilities in GPT‑5.6, not a sudden power grab, and a sign that frontier models are now treated more like sensitive dual‑use technologies than ordinary software. Critics inside the tech policy world warn that, without clear safety standards, the emerging de facto licensing regime around frontier AI could lead to unpredictable launch delays and weaken U.S. competitiveness versus China.

Online, the restrictions are being interpreted against a longer history of “too dangerous to release” debates. One widely shared post recalled how an OpenAI leader once delayed the public release of GPT‑2 on safety grounds, suggesting that political pressure is now amplifying a long‑running tension between caution and openness in AI development.

Continue reading https://foxvector.com/stories/019f082e-fdaa-1309-732d-2f3bcfaf3451