OpenAI is staggering the release of its latest AI model, GPT 5.6, following a request from the US government, mirroring the launch strategy of Anthropic's Mythos. CEO Sam Altman informed staff this week that GPT 5.6 would initially be available only to a small group of partners, as reported by The Information.
Government Approval Required
Altman stated in a memo that the federal government would be "approving access customer by customer during this preview period" for GPT 5.6. If the process proceeds smoothly, the model will be released more broadly "a couple of weeks later." The CEO emphasized that this approach is not OpenAI's preferred long-term model and that the company will work with the government and industry to develop a more sustainable release strategy for future models.
Echoes of Anthropic's Mythos
Anthropic, a close rival of OpenAI, previously employed a similar release program for its Mythos model. However, Anthropic has since withdrawn the technology entirely after the US government ordered it to prevent foreign nationals from accessing public versions of the model, citing its powerful cyber-hacking capabilities.
Government Agencies Involved
OpenAI had been collaborating with two government agencies—the Office of the National Cyber Director and the Office of Science and Technology Policy—on the preview. According to The Information, Commerce Secretary Howard Lutnick intervened against even a limited release, calling Altman to demand approvals from additional agencies.
Shift in White House Stance
This month, President Donald Trump signed an executive order creating a voluntary framework for the federal government to vet powerful new AI models before release. This marks a departure from the White House's previous deregulatory approach; Vice President JD Vance had warned last year that "excessive regulation of the AI sector could kill a transformative industry." The policy shift comes amid rapidly improving model capabilities, including Mythos, which the UK's AI security body described as a "step up" over previous cutting-edge models.
OpenAI has been approached for comment but has not yet responded.
