OpenAI limits GPT‑5.6 Sol to trusted partners after US government

OpenAI announced that its newest GPT‑5.6 Sol model is available only to a “small group of trusted partners” after a request from the U.S. government, prompting users to warn that a divide between AI haves and have‑nots has begun.

Government‑driven restriction

OpenAI said the restriction is a temporary step while the White House conducts a cybersecurity review of the Sol model, the flagship of the GPT‑5.6 series [1]. The request comes under an executive order signed by President Trump in June that allows the federal government to vet advanced AI systems for up to 30 days before public release [2]. OpenAI’s statement emphasized that it “doesn’t believe this kind of government access process should become the long‑term default,” framing the limitation as a short‑term safeguard rather than a new norm [1].

User reaction and market implications

The limited rollout has sparked frustration on platforms like Reddit, where users describe the situation as the start of an “AI divide” between those with privileged access and the broader public [1]. Competitors are watching closely; Anthropic recently withdrew two models (Fable 5 and Mythos 5) after a similar government directive, highlighting a growing trend of regulatory pressure on frontier AI labs [2][3]. OpenAI claims Sol matches Anthropic’s Mythos model on cybersecurity tasks while using a third of the output tokens, but these performance assertions remain unverified by independent benchmarks [1].

Competitive context

OpenAI positions the GPT‑5.6 series as a step up from GPT‑5.5, citing improvements in affordability, safety, coding, biology, and cybersecurity [1]. The company also notes that all three models—Sol, Terra (for everyday work), and Luna (the smallest and fastest)—are built with stronger safeguards against adversarial prompts [1]. However, the lack of public testing means developers and enterprises cannot yet assess whether the claimed safety gains translate into real‑world advantages over rivals like Anthropic or other large‑language‑model providers.

What to watch

• Rollout timeline – OpenAI has said broader availability could arrive “in the coming weeks,” making the exact release date a key milestone.
• Regulatory developments – Further actions under the June executive order could shape how quickly advanced models reach the market.
• Competitor responses – Anthropic’s next move after its model withdrawals may signal whether other labs will adopt similar restricted‑release strategies.

The restriction underscores a tension between rapid AI innovation and emerging national‑security oversight, leaving the industry to gauge how government vetting will shape access to the most capable models.