The claim rests on a single social media post by journalist Stephanie Palazzolo of The Information, shared the same day as the current date. Her quoted message says the Trump administration has asked OpenAI to stagger the release of GPT-5.6 over security concerns, with CEO Sam Altman telling staff that the government would approve access "customer by customer." A separate post by Rohan Paul repeats this framing and adds that the concern centers on automated higher-skill cyber work, where the same model could help both defenders and attackers. The actual The Information article is not available in the materials provided, so the post is the sole direct evidence for the central claim.
OpenAI's own publications tell a related but different story. On June 18, 2026, the company announced GPT-5.5, describing it as already rolling out to ChatGPT and Codex users, with API access still being worked through with partners and customers. Three days later, on June 22, 2026, OpenAI published a separate post introducing GPT-5.5-Cyber and explaining its Trusted Access for Cyber (TAC) framework, an identity-based system that restricts the model's most permissive cybersecurity capabilities to verified defenders. OpenAI's system card for GPT-5.5 reinforces that the company added targeted testing for advanced cybersecurity capabilities before release.
These sources show OpenAI has indeed built a security-gated access tier for its cyber-capable models, but they refer specifically to GPT-5.5, not GPT-5.6, and they describe a company-led framework rather than a government-directed rollout. Nothing in the provided materials mentions GPT-5.6, a Trump administration request, or customer-by-customer government approval. Because the central claim concerns a model version that does not appear in any verifiable primary source, and the only supporting material is an unverified excerpt of paywalled reporting, the claim remains unverified.
