OpenAI Restricts GPT-5.6 Rollout at US Government's Request

Jakarta - Artificial intelligence company OpenAI is restricting the circulation of its latest product, GPT-5.6, due to a request from the United States government. The request comes as the US government applies new pressure on AI companies to limit their most advanced systems, as reported by TechCrunch on Friday (26/6). President Donald Trump issued an executive order requesting that several AI companies voluntarily report their most advanced AI models for government review up to 30 days before release. As a result, the preview of GPT-5.6 is currently limited to partners ‘whose participation has been notified to the government’. OpenAI is not the only company facing such restrictions. After Anthropic released its most advanced public model, Fable 5, the US ordered the company to remove access for any foreign nationals, prompting Anthropic to remove the model entirely. The incident has raised questions about how much power the government should have over AI model releases. In response, OpenAI stated it was unhappy with the rule. Despite this, the company complied with the government’s request. ‘We do not believe the government access process should be a permanent setting. It limits the best tools from being accessed by users, developers, companies, cyber defence, and global partners who need them,’ OpenAI said in a post on its official blog. Former White House AI advisor and prospective OpenAI employee Dean Ball said that President Trump’s executive order has created a de facto forced licensing regime for cutting-edge AI, leading to excessive restrictions. According to Ball, the issue is further complicated by the government’s lack of clear safety standards, which could cause launch delays, benefit China in the AI race, and jeopardise billions of dollars allocated for AI infrastructure. GPT-5.6 is the company’s most powerful model to date, featuring enhanced agent capabilities in coding, biology, and cyber security. Sol introduced a maximum reasoning effort mode and an ultra mode that uses coordinated sub-agents to complete highly complex tasks.