OpenAI said it will publicly release its highly advanced GPT-5.6 artificial intelligence (AI) model on Thursday, ending a two-week standoff with the Trump administration over national security and regulatory oversight.
Wide release of the flagship model, named Sol, alongside its lower-tier counterparts, Terra and Luna, comes after OpenAI initially bowed to White House pressure last month to restrict access to a small group of government-approved partners. The delay was prompted by federal concerns regarding potential cybersecurity, biological, and coding risks inherent to the advanced technology.
Described by OpenAI as its “strongest model yet,” GPT-5.6 Sol will become widely available to the public, enterprises, and global developers starting Thursday.
“Happy building,” OpenAI CEO Sam Altman wrote in a brief post on social media platform X late Tuesday, signaling the conclusion of additional safety testing conducted by the Department of Commerce’s Center for AI Standards and Innovation.
Friction between Silicon Valley and Washington highlights an increasingly turbulent relationship as the government attempts to navigate AI regulation in real time.
Last month, President Trump signed a sweeping AI executive order establishing a voluntary framework that allows federal agencies to test cutting-edge models for up to 30 days prior to their public debut.
While the administration has maintained that these evaluations are strictly voluntary, critics and industry insiders argue the White House is enforcing an “ad hoc” pre-clearance system. OpenAI openly chafed at the initial deployment restrictions last month, warning that a staggered rollout “keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them.” The company added that such government intervention should not become the long-term default.
Compounding the industry’s anxiety, a White House official actively disputed the notion that the administration had “cleared” or “green-lit” OpenAI’s launch, telling reporters that no such federal permission is legally required. The official stressed that under the June 2 executive order, mandatory licensing is explicitly barred, and the timing of any release “rests entirely with the companies.”
Yet, OpenAI is not the only tech giant to face off with the administration.
Rival Anthropic was recently forced to abruptly pull its advanced Mythos and Fable models from the market following a sharp export control directive from the Commerce Department that banned foreign access. Though the restrictions on Anthropic were lifted last week, the aggressive federal interventions have sparked widespread confusion and backlash among AI policy advocates, who warn that unpredictable government overreach could stifle American innovation.
On Wednesday, Anthropic extended its promotional access window for its flagship model, Claude Fable 5, pushing the expiration date to July 12. According to an update on the company’s official support page, the extension allows eligible subscribers to continue leveraging the model’s advanced capabilities as part of their existing tiers. During this promotional window, users can allocate up to 50% of their weekly subscription limits to Claude Fable 5 at no additional cost. Once this threshold is reached, users retain the flexibility to either burn usage credits to maintain access or switch to alternative Claude models to preserve their remaining plan limits.
The extension comes at a critical juncture for the generative AI sector, widely seen by industry analysts as a strategic play to lock down developer mindshare. By lowering the barrier to entry for its premier model, Anthropic aims to anchor its user base ahead of major shifts in the ecosystem, especially as GPT-5.6 comes to market Thursday.
For now, OpenAI’s global rollout is proceeding as planned. The company said it is using the GPT-5.6 launch to help the administration establish a more predictable, repeatable framework for future model releases.
