OpenAI restricts limited release of new model to US only

OpenAI has initiated a restricted preview of its newest advanced AI model series, exclusively for US-based partners, a move prompted by a request from the US government. This decision comes two weeks after the White House directed OpenAI's competitor, Anthropic, to block foreign nationals from accessing its powerful AI models, citing national security risks.

Anthropic complied by shutting down access to its Fable 5 and Mythos 5 models, stating it could not effectively enforce the restriction on foreign nationals. The latest AI models from leading companies like Anthropic and OpenAI are raising significant concerns due to their advanced capabilities, particularly in identifying software vulnerabilities that could be exploited by hackers. In response to these novel abilities, President Trump signed an executive order earlier this month establishing a voluntary federal review of national security risks associated with advanced AI models before their public release.

“

We don’t believe this kind of government access process should become the long-term default.

The White House has provided limited details on how it plans to enforce this executive order, which relies on voluntary participation from companies. This intervention is notable given the administration's broader push to reduce AI oversight and its efforts to prevent states from enacting their own regulations. The strong action against Anthropic has led to accusations of government overreach.

“

It keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them.

OpenAI acknowledged its discomfort with the required government access process for its new models. The company stated it briefed the US government on the capabilities of its new models beforehand and, at the government's behest, is commencing with a limited preview for a select group of trusted partners whose identities have been vetted by authorities. While these partners are US-based, OpenAI indicated that overseas employees of these companies would also have access.

In a blog post, OpenAI expressed that this type of government access process should not become the long-term norm, as it restricts access to essential tools for users, developers, and global partners. However, the company views this short-term measure as the most effective route to achieving broader availability in the coming weeks. The situation echoes earlier tensions where Anthropic reportedly angered the Trump administration by refusing to allow its technology for mass surveillance and autonomous weapons, leading to the cancellation of Pentagon contracts.

“

We are taking this short-term step because we believe it is the strongest path to broader availability in the coming weeks.