OpenAI is set to publicly release GPT-5.6 this Thursday, July 9, following approval from the Trump administration. The company announced the wider rollout on X, confirming that all three variants of the new model series will be available globally. “We’re expanding preview access globally now,” the company said.
The path to this public release was not straightforward. GPT-5.6 was first made available in late June, but only to a small group of trusted partners. That limited rollout was a direct result of an AI cybersecurity order President Trump signed in early June, which asks AI companies to voluntarily submit their most powerful models for government review 30 days before any public release.
OpenAI was not entirely happy about the arrangement. The company said at the time that it did not think “this kind of government access process should become the long-term default,” but chose to comply because it was the fastest way to get the model out to the public. As it turned out, the wait was shorter than 30 days.
According to Axios, the Trump administration gave OpenAI the go-ahead for a broader release after additional testing and more meetings between the company and officials. The Department of Commerce’s Center for AI Standards and Innovation ran those tests. OpenAI also sent technical experts to Washington, D.C. to answer questions and address any concerns in real time.
GPT-5.6 comes in three distinct variants, each aimed at a different use case:
- Sol is the company’s most capable model to date, aimed at complex, demanding tasks
- Terra is built for everyday use, with performance comparable to GPT-5.5 but at half the cost
- Luna is the lowest-cost option, designed for high-volume, budget-conscious applications
Pricing reflects that tiered structure. Sol costs $5 per million input tokens and $30 per million output tokens. Terra comes in at $2.50 per million input and $15 per million output. Luna is the cheapest at $1 per million input and $6 per million output.
OpenAI is not the only AI company navigating this new government review process. Anthropic was ordered to block all foreign nationals from accessing its latest Mythos cybersecurity and Fable models to stay compliant. The company has since received approval to redeploy Mythos 5, with Fable 5, its broader public-facing counterpart, expected to follow.
The pattern here is worth watching. Whether this review process becomes a standard fixture for major AI releases, or remains an informal arrangement that companies tolerate rather than welcome, will likely shape how quickly frontier models reach users going forward.
