Trump Administration Presses OpenAI to Limit GPT-5.6 Release to Select Partners

After Anthropic Showdown, Trump Administration Presses OpenAI To Limit GPT-5.6 Release: Report

The Trump administration has reportedly asked OpenAI to limit the distribution of its forthcoming model, GPT-5.6, to a select group of government-approved partners, citing potential security risks. In the memo, Altman said that they had made it clear to the U.S. government that this was not their preferred long-term model and that they would work with the government and others in the industry to achieve a more sustainable approach for future releases. Another report by Axios indicated that Altman spoke with Commerce Secretary Howard Lutnick about GPT-5.6 on Wednesday. The publication reported that the government's intervention was driven by GPT-5.6's "Mythos-like" capabilities, rather than any broader shift toward increased government oversight. A source said the administration is focused on ensuring companies implement adequate safeguards for highly advanced AI models of that caliber. OpenAI and the White House did not immediately respond to Benzinga's request for comments. Trump's AI Crackdown Expands The move follows growing concerns about advanced AI misuse and the Trump administration's clash with Anthropic, which disabled access to its Fable 5 and Mythos 5 models after being directed to block foreign nationals from using them earlier this month. Later, President Donald Trump said he initially viewed Anthropic as a potential national security risk but now believes the AI startup acted responsibly after complying with a U.S. directive to restrict access to its most advanced AI models. Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors. Image via Shutterstock Market News and Data brought to you by Benzinga APIs To add Benzinga News as your preferred source on Google, click here.

Like Anthropic, OpenAI may restrict GPT 5.6 access during initial rollout: Here is why

OpenAI CEO Sam Altman reportedly told employees that the government would be "approving access customer by customer" during the preview period. OpenAI may take a different approach with the launch of its upcoming AI model, GPT-5.6. Instead of making the model available to everyone, the company will reportedly give access only to a select group of close partners. According to a report by The Information, this is because the Trump administration has asked the AI company to do so. At a meeting this week, OpenAI CEO Sam Altman reportedly told employees that the government would be "approving access customer by customer" during the preview period. He also reportedly said that if the early rollout goes smoothly, OpenAI hopes to expand access with a broader public release a "couple of weeks later." The report says OpenAI has been working closely with the US government ahead of the release. The agencies that asked for the limited rollout include the Office of the National Cyber Director and the Office of Science and Technology Policy. Also read: Google reportedly postpones Gemini 3.5 Pro launch, here is why The move comes as the Trump administration has increased its focus on advanced AI systems. The administration had earlier supported a more hands off approach to AI regulation. But now, it has recently pushed for federal oversight. Earlier this month, President Donald Trump signed an executive order asking certain AI companies to voluntarily submit new AI models to the government for testing and evaluation before making them publicly available. If OpenAI follows this plan, it would be taking a path similar to Anthropic. Earlier this year, Anthropic introduced its cyber-focused AI model, Claude Mythos, but limited access to a small group of partners through a program called Project Glasswing. The company said the model was too powerful to release widely because it could be misused. Also read: GTA 6 pre orders now live: India price, benefits and other details One of the biggest concerns around powerful AI models is their ability to help with cyberattacks. Models designed for cybersecurity could become even more powerful. They may be able to find and exploit software vulnerabilities much faster than human researchers. Meanwhile, OpenAI announced a new version of GPT-5.5 Instant, the AI model that powers ChatGPT by default, earlier this week. The company said the latest version is smarter, better at understanding users, and more fun to talk to. The company also introduced Jalapeno, its first AI chip built for LLMs workloads.