OpenAI to limit GPT-5.6 rollout after federal request
The Trump administration will reportedly review which customers can use GPT-5.6 during a limited preview tied to security concerns.
By Maya Lindqvist · Senior Technology Correspondent
2 min read
OpenAI will hold back broad access to GPT-5.6 after the Trump administration asked the company to phase in the model’s release, according to The Information. The reported arrangement matters because it puts federal officials in the middle of customer access decisions for one of the next major AI systems from a leading U.S. developer.
The Information reported that OpenAI CEO Sam Altman told employees during a company Q&A on Wednesday that GPT-5.6 would launch first as a limited preview. Under that plan, only a small group of enterprise customers would get access at the start, according to the report.
The Information reported that OpenAI is taking that approach to comply with a federal request driven by concerns about possible security risks. During the preview period, the Trump administration would approve customer access individually, according to the report.
The reported OpenAI plan differs from how the administration recently handled Anthropic, a rival AI company, according to The Verge. Earlier this month, the administration gave Anthropic an ultimatum requiring it to suspend access to its Mythos 5 and Fable 5 models, The Verge reported.
The Verge reported that the administration used an export control directive in Anthropic’s case. That directive barred “foreign nationals” from using the technology, including Anthropic employees who were not U.S. citizens, according to The Verge.
Federal AI controls draw industry concern
The Trump administration’s interventions come after it had signaled a faster approach to artificial intelligence policy. The Verge cited a Department of War AI strategy that used the phrase “speed wins,” along with a Commerce Department effort to encourage an American AI exports program.
The Anthropic restrictions raised concern across the technology industry, according to The Verge. The reported GPT-5.6 arrangement suggests the administration is applying controls differently by company, with OpenAI allowed to proceed through a limited preview while Anthropic faced a suspension order, The Verge reported.
The Information did not report a date for wider GPT-5.6 availability. It also did not report which enterprise customers would receive access first or what standards the government would use when deciding whether to approve customers during the preview.
OpenAI has not announced the GPT-5.6 release plan publicly in the reported terms. The Information’s account centers on Altman’s internal remarks to employees and on the federal request that shaped the rollout.
This story draws on original reporting from The Verge.