OpenAI gets Nod from Trump Admin to Release its Most Powerful Model

Mumbai, July 8: OpenAI appears to have gotten Washington out of its way. The Trump administration has quietly given OpenAI the green light to publicly release its most powerful language model yet, GPT-5.6, Axios reported Tuesday. The approval came after weeks of additional safety reviews and meetings with government officials, according to the report.

The news was announced Tuesday night when OpenAI said its flagship AI language model, dubbed Sol, as well as two smaller versions, Terra and Luna, would be released for public use Thursday. As with OpenAI’s last major release, GPT-5 last month, getting to this point was more telling about Washington’s relationship with the world’s most powerful AI companies than the technology itself.

Rather than simply reviewing OpenAI’s safety testing ahead of time as the company originally wanted to do with GPT-5, officials from the Trump administration have been negotiating access and release terms with AI companies on what seems like a case by case basis and in real time. It’s akin to some bizarre diplomatic dance nobody has the steps for just yet.

That back and forth has required OpenAI, which is based in San Francisco, to essentially camp out in Washington as the administration tries to assess potentially harmful uses for its technology. The additional sign-off came from the Center for AI Standards and Innovation, an office within the Department of Commerce, according to Axios.

This is the second such incident the administration has inflicted on OpenAI this summer. In June, officials strongly encouraged OpenAI to use a staggered rollout for GTP-5.6, which gave governments and their approved entities early access to the model. OpenAI was unequivocal at the time that it did not prefer that model of release.

OpenAI executives have said this entire process of AI approvals feels improvised. Both sides are effectively operating on regulations and standards that have yet to be clearly defined, even after Trump’s May executive order on AI urged the creation of clearer rules. Many of those guidelines have still not been formally proposed despite the July deadline.

The company is hardly alone in this experience. Earlier this summer, the Commerce Department essentially removed Anthropic’s popular chatbot models Mythos and Fable from the international market, blocking foreign customers from accessing the technology in June.

The department later lifted its restrictions on Fable last week, with full customer access restored the day after. The about-face was relatively swift for a federal agency’s standards.

Both incidents send a clear signal. If recent events are any indication, the release of advanced AI models will be dictated by Washington’s timetable as much as Silicon Valley’s moving forward.

Source: https://www.axios.com/2026/07/08/openai-gpt-trump-ban-lifted