OpenAI Introduces Its Advanced GPT-5.6 and Additional Models After US Government Approval
OpenAI confirmed that its latest artificial intelligence models, called GPT-5.6 Sol, Terra and Luna, will be made available to the public this Thursday.
Prior access to the GPT-5.6 version had been restricted to a select group of partners, following a delay request made by the US government last month.
Recently, information released indicated that the Trump administration granted authorization for the large-scale launch, after carrying out additional assessments.
OpenAI is set to launch GPT-5.6, its most sophisticated AI model, this Thursday, after a month-long delay caused by requests from the United States government, which has expressed growing concerns about national security and the potential misuse of high-powered AI technologies.
More on this story: Google introduces Gemini 3.1 Flash-Lite, an AI model for fast, cost-effective responses
Currently, the United States and China are in a fierce dispute to develop cutting-edge artificial intelligence models, which, in the opinion of experts, could significantly accelerate complex cyber attacks in sectors dependent on old, interconnected and intricate technological systems.
In Washington, there has been increased scrutiny over the release of advanced AI models. The objective is to identify possible threats, amid fears that this technology could be used inappropriately by military forces or intelligence services in countries such as China and Russia, among others. This scenario represents a milestone in government intervention in technological development, marking an era in which national security imposes an unprecedented level of scrutiny on innovation.
Learn more: New Claude Sonnet 5 version arrives with advanced autonomy and competitive cost-benefit in the AI market
Chinese authorities have also held meetings with major technology companies to discuss the possibility of imposing restrictions on foreign access to China’s most advanced AI models, including those that have not yet been launched on the market.
Anthropic, one of OpenAI’s main competitors, unexpectedly suspended all users’ access to its most advanced AI models, Mythos 5 and Fable 5, on June 12. The measure was taken following an export control order issued by the US government, due to national security concerns. The restrictions were later lifted last week after the company implemented new safeguards.
As reported, the Trump administration has given the green light for the large-scale launch of GPT-5.6. This approval came after a series of additional tests and meetings between company representatives and government authorities.
See Also
Munich: Boeing 787 Vietnam Airlines scrapes its tail on takeoff and makes an emergency landing after taking off off the runway
Advance of Cyclone Lala paralyzes airports and raises alert in Hawaii
Hurricane Lala arrives in Hawaii with winds of 120 km/h and closes airports
Both the White House and the United States Department of Commerce were silent on the matter, not responding to requests for comment outside of regular business hours.
Initial access to GPT-5.6 was granted to a limited number of partners selected by OpenAI, and data from these users was shared with relevant authorities.
On the same topic: Google launches Gemini 3.5 with cutting-edge intelligence and speed for automation
The company that created ChatGPT confirmed that it will make available the GPT-5.6 Sol, its most advanced model, in addition to the Terra and Luna models, which will have a more affordable cost. The announcement was made in a publication on the Google+ platform last Tuesday night.
When introducing the new models in June, OpenAI highlighted its enhanced capabilities in areas such as programming, biology and cybersecurity. The company highlighted that GPT-5.6 Sol was highly competitive with Anthropic’s Mythos Preview in cybersecurity tests such as ExploitBench.
This Wednesday, billionaire Elon Musk, through his company SpaceXAI, also announced the public availability of his main artificial intelligence model, Grok 4.5, intensifying the rivalry with Anthropic and OpenAI.
Understanding National Security Concerns
US President Donald Trump formalized an executive order establishing a voluntary framework. Through it, artificial intelligence developers can deliver “covered frontier models” to the US government for a period of up to 30 days, before releasing them to partners deemed trustworthy.
More on this story: Scarcity of training data threatens to limit the advancement of artificial intelligence soon
Although the government in Washington has lifted export controls for Anthropic’s Fable model, Mythos, which was developed specifically for cybersecurity professionals, is still only available to a select number of “trusted” organizations in the United States.
In China, officials express concern about Mythos’ potential to exploit software vulnerabilities, as well as concerns that the US could use the model against Beijing’s interests.
Anthropic even issued a warning, stating that it would be “probably impossible” to develop any AI model that was completely immune to system intrusions.
