OpenAI, Anthropic Launch AI Models Under US Government Gated Access

OpenAI released GPT-5.6 Sol, Terra, and Luna in a limited preview restricted to ~20 organizations approved by the Trump administration, marking the first government-gated frontier AI rollout. Simultaneously, the US Commerce Department partially lifted restrictions on Anthropic's Mythos 5, allowing over 100 trusted US organizations to access the model after a two-week standoff. The moves signal a new era of federal oversight of advanced AI capabilities, while Google continues to lose top AI researchers to Anthropic and OpenAI.

Written by AI from the sources it cites; it can be wrong — follow any citation to confirm.

14 sources1 min read

OpenAI debuts GPT-5.6 Sol under first government-gated AI release in US history OpenAI launched its GPT-5.6 model suite — including flagship Sol, mid-tier Terra, and fast Luna — on June 27, but limited initial access to roughly 20 organizations individually approved by the Trump administration. The rollout follows a White House executive order requiring federal vetting of frontier models with advanced cyber capabilities, and OpenAI signaled it does not want this government access process to become the long-term default. 1 2 3 4 5 6

US government partially lifts restrictions on Anthropic's Mythos 5, allows limited rollout The Commerce Department granted Anthropic permission to restore access to its Mythos 5 cybersecurity model to over 100 trusted US companies and federal agencies, after a two-week standoff over national security concerns. The deal requires Anthropic to implement safeguards, while Fable 5 remains offline pending further negotiations; the administration's interest in managing AI model releases appears to be expanding. 5 6 7 8 9 10 11

Google loses top Gemini researchers to Anthropic as AI talent war intensifies Jonas Adler and Alexander Pritzel, key contributors to Google's Gemini model, are leaving for Anthropic, following recent high-profile exits by Noam Shazeer (to OpenAI) and Nobel laureate John Jumper (to Anthropic). The departures raise concerns about Alphabet's ability to retain AI talent as rivals prepare for public offerings and offer significant equity incentives. 12 13 14

Commerce's partial lift on Mythos 5 exposes legal gaps in AI export control framework The Commerce Department's decision to exempt 'certain trusted partners' from export license requirements for Anthropic's Mythos 5 highlights a fundamental legal uncertainty: the Export Administration Regulations were designed for physical goods, not cloud-accessed AI capabilities. Experts argue that the 'deemed export' doctrine may not cleanly map onto global API access, underscoring the need for layered defenses combining technical safeguards, monitoring, and government oversight. 10