OpenAI slows GPT-5.6 rollout after US government request—and pushes back publicly

OpenAI announced last week that it will roll out its newest AI model family—GPT-5.6—to only a small group of pre-approved partners, rather than releasing it openly as planned. The decision came after a direct request from the U.S. government's Office of the National Cyber Director and Office of Science and Technology Policy. This marks the first time the U.S. government has asked an American AI company to restrict a model launch before release. OpenAI CEO Sam Altman also received a call from Commerce Secretary Howard Lutnick, warning the company to hold back pending government sign-off from additional agencies. The GPT-5.6 family includes three models: Sol (the flagship), Terra (for balanced, everyday use), and Luna (the faster, lower-cost option). Sol features improved capabilities in coding, biology, and cybersecurity, plus new reasoning modes designed for highly complex tasks.

According to sources, the government intervened because GPT-5.6 Sol has what one official described as exceptional capability—comparable to previous advanced models that drew regulatory attention. The administration is trying to build a framework for testing and evaluating the security risks of cutting-edge AI systems before they reach the broader market. This request sits within a larger pattern: Anthropic faced similar pressure to pull its Fable 5 and Mythos 5 models offline, and the Trump administration issued an AI executive order in June designed to create a voluntary (though increasingly mandatory-sounding) government review process for new releases.

While OpenAI complied with the government's request, the company made its position clear in a public blog post on Friday. OpenAI stated: "We don't believe this kind of government access process should become the long-term default. It keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them." Sam Altman sent an internal memo to staff acknowledging the decision but emphasizing: "We've made clear to the U.S. government that this is not our preferred long term model, and will work with them and others in industry to achieve a more sustainable approach for future releases." The company framed the preview as a short-term step, saying GPT-5.6 will move toward broader availability within a couple of weeks as it collaborates with the administration on a new cybersecurity framework.

For Australian mid-market firms, the practical impact is clear: world-class AI tools are arriving slower than they could be, and access depends on government approval. A retailer hoping to deploy advanced AI for inventory forecasting, a professional services firm seeking better client analytics, or a healthcare provider looking to streamline operations all face delays and uncertainty. The regulatory pattern emerging in the U.S. is likely to influence how AI tools roll out globally, including here. If approval processes become standard, the businesses most likely to benefit—smaller and mid-sized companies without lobbying power—may be the last to access them.

As AI regulation and model access become more complex and fragmented, the foundation underneath matters more than ever. Businesses that have already unified their data—pulling together insights from their CRM, accounting system, customer records, and operations—are best positioned to extract value from whatever AI tools they can access, whenever approval comes through. Rather than waiting months to get a new AI service running, companies with a clean, centralized data foundation can deploy new capabilities in weeks. They're also better prepared to comply with government requests for audits or testing, since their data landscape is transparent and governed. Right now, the ability to move fast with AI isn't about having the fanciest model; it's about having your data ready.