OpenAI is preparing to publicly release its most advanced model, GPT-5.6, on Thursday after delaying the launch last month at the request of the U.S. government.
The delay came amid rising national security concerns that powerful artificial intelligence systems could be misused for cyberattacks, military purposes or intelligence operations.
U.S. and China Compete in Advanced AI Race
The United States and China are both racing to develop the most advanced AI models.
Experts have warned that these systems could significantly speed up complex cyberattacks, especially against sectors that depend on interconnected and outdated technology infrastructure.
Because of these risks, Washington has increased scrutiny of major AI model releases.
U.S. Government Reviews Frontier AI Models
U.S. officials are concerned that advanced AI technology could be misused by military or intelligence actors in China, Russia or other countries.
As a result, the U.S. government has been reviewing powerful model launches more closely to identify potential national security threats before wider release.
Chinese authorities have also reportedly held discussions with leading technology companies about possibly restricting overseas access to China’s most powerful AI models, including models that have not yet been released.
Anthropic Models Previously Restricted
OpenAI’s rival Anthropic recently disabled access to its most advanced AI models, Mythos 5 and Fable 5, for all users.
That move followed a June 12 U.S. export control order issued over national security concerns.
The restrictions were partly lifted last week after Anthropic added certain safeguards.
However, while export controls were lifted for the Fable model, Mythos, which is designed for cybersecurity professionals, remains available only to some trusted U.S. organizations.
Trump Administration Approves GPT-5.6 Launch
According to Axios, which first reported the news, the Trump administration approved a broad public launch of GPT-5.6 after additional testing and meetings between OpenAI and government officials.
The White House and the U.S. Department of Commerce did not respond to Reuters’ request for comment outside regular business hours.
OpenAI Had Limited Early Access
Before the public release, OpenAI limited access to GPT-5.6 to a small group of vetted partners.
Details about those partners were reportedly shared with government authorities.
OpenAI said in a post on X late Tuesday that it will launch its most capable model, GPT-5.6 Sol, along with lower-cost models called Terra and Luna.
Elon Musk Announces Grok 4.5 Public Release
Billionaire Elon Musk also said Wednesday that his company, SpaceXAI, is making its leading AI model, Grok 4.5, available to the public.
Musk’s company competes with both OpenAI and Anthropic in the advanced AI market.
Executive Order Creates Voluntary AI Review Framework
U.S. President Donald Trump has signed an executive order creating a voluntary system for AI developers.
Under the framework, companies can provide “covered frontier models” to the U.S. government for up to 30 days before releasing them to trusted partners.
The goal is to allow officials to review powerful AI systems before they become widely available.
China Raises Concerns About Cybersecurity Model
In China, officials are reportedly concerned about Anthropic’s Mythos model.
Authorities worry the model could be used to exploit software vulnerabilities or be deployed by the U.S. against Chinese interests.
Anthropic has also warned that it is probably impossible to make any AI model completely resistant to jailbreaks.
OpenAI’s planned public release of GPT-5.6 comes at a time when advanced AI models are under growing national security scrutiny. The U.S. government delayed the launch to conduct further testing and review possible risks, while similar concerns have already affected Anthropic’s most powerful models.
With the United States, China and major AI companies competing to release frontier systems, governments are increasingly focused on how to balance innovation, cybersecurity and global security risks.