Topline<\/p>\n
OpenAI said Friday it is rolling out three new AI models in the company\u2019s GPT-5.6 series on Friday\u2014but the advanced technology is only available to select \u201ctrusted partners\u201d at the request of the U.S. government ahead of a larger rollout in the coming weeks.<\/p>\n
OpenAI’s GPT-5.6 technology is only available to select users approved by the U.S. government. (Photo by Nathan Posner\/Anadolu via Getty Images)<\/p>\n
Anadolu via Getty ImagesKey Facts<\/p>\n
OpenAI on Friday announced<\/a> the rollout of Sol, the flagship GPT-5.6 model, alongside Terra, a model for daily work, and Luna, a more affordable version.<\/p>\n The company said a broader rollout will follow in the coming weeks, but said it is starting with a \u201climited preview for a small group of trusted partners\u201d approved by the U.S. government.<\/p>\n Multiple<\/a> news outlets reported<\/a> Thursday the Trump administration had requested OpenAI limit the release of its GPT-5.6 model to government-approved users due to concerns about the models\u2019 advanced capabilities. <\/p>\n OpenAI said in its Friday announcement that government approval for its AI models should not be the \u201clong-term default,\u201d saying it \u201ckeeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them.\u201d<\/p>\n The GPT-5.6 rollout comes weeks after President Donald Trump signed an executive order<\/a> that asked AI companies releasing advanced models to first share the technology with the government to assess their capabilities and select partners who can access the technology before a wider release. <\/p>\n What Do We Know About The New OpenAI Models?<\/p>\n OpenAI described<\/a> Sol as its \u201cstrongest model yet.\u201d The company said the model shows advanced capabilities in \u201ccoding, biology, and cybersecurity,\u201d and it claimed the Sol model outperformed Anthropic\u2019s powerful Mythos technology on Terminal\u2011Bench 2.1, an AI benchmark that measures how well a model can complete complex tasks. The company also said the new models were built with its \u201cmost robust safeguards to date,\u201d also claiming Sol is its \u201cmost capable model yet for cybersecurity.\u201d<\/p>\n How Has The Trump Administration Increasingly Regulated AI?<\/p>\n Weeks after signing the executive order earlier this month requesting early access to powerful AI models, the Trump administration reportedly directed<\/a> Anthropic to disable access for all foreign nationals to its powerful Mythos models, citing national security concerns. In response, Anthropic disabled access to those models for all customers, saying in a post<\/a> on X the move was necessary to ensure full compliance. The New York Times<\/a> reported last week some Anthropic employees and cybersecurity experts believe the Trump administration is unfairly targeting the company. Anthropic and the Trump administration faced off earlier this year when the company\u2019s CEO Dario Amodei objected to the Defense Department using Anthropic technology for mass domestic surveillance and to power autonomous weapons. Secretary of Defense Pete Hegseth, in response, designated the company a supply chain risk to national security, and in March, Anthropic sued the administration alleging it illegally retaliated against the company for its position on AI safety. <\/p>\n what to watch for<\/p>\n OpenAI is preparing to go public, but reports earlier this week indicate it may delay<\/a> its initial public offering until 2027. OpenAI is one of a trio of companies\u2014alongside SpaceX, which went public earlier this month, and Anthropic, which is also planning to go public\u2014planning a blockbuster IPO. The New York Times<\/a> reported SpaceX\u2019s stock price, which has slid since its debut on the public market, was a factor that influenced OpenAI to delay its IPO plans. <\/p>\n further reading<\/p>\n