OpenAI announces global release of GPT-5.6 models after US lifts restrictions

OpenAI is set to release its most advanced AI model, GPT-5.6, globally on Thursday, July 9. This move comes after the US lifted restrictions on access to GPT-5.6 which had been placed last month. GPT-5.6 has three models – Sol, Terra, and Luna.

In Short

• OpenAI announces global release of GPT-5.6
• US had previously asked for limited GPT-5.6 access due to cybersecurity concerns
• GPT-5.6 is OpenAI's most advanced AI model yet

OpenAI is finally bringing its most advanced AI models to the public. After initially restricting the GPT-5.6 family of AI models to select companies, the AI startup has announced that GPT-5.6 will be launched globally on Thursday, July 9. This comes after the US government lifted restrictions on access to GPT-5.6.

On X, OpenAI confirmed the public release for GPT-5.6 on X. The AI startup wrote, “GPT-5.6 Sol, along with Terra and Luna, will launch publicly this Thursday. We’re expanding preview access globally now."

OpenAI CEO Sam Altman tweeted, "GPT-5.6 sol launches thursday! happy building."

As part of the initial release of GPT-5.6 late last month, the AI models were only accessible to a select group of organisations. This was done after the US government asked OpenAI for a limited release over concerns of cybersecurity.

You see, OpenAI says that GPT-5.6 Sol is its most advanced model yet. This could potentially make it easier for bad actors to find and exploit cybersecurity flaws online – something that was also a concern with Anthropic’s Claude Mythos.

Now, according to a report from Axios, the US Department of Commerce approved a broad launch of GPT-5.6 after additional testing under Washington’s new oversight framework for frontier AI.

What is GPT-5.6?

GPT-5.6 has three models – Sol, Terra, and Luna. Sol is said to be the most advanced and most expensive AI model from OpenAI, aimed at complex tasks including coding, research, reasoning and solving difficult problems.

The AI startup also says that Sol shows gains in coding and biology and is its strongest model so far for cybersecurity-related tasks, including helping fix software security issues, while remaining within the company’s internal safety limits for high-risk behaviour.

On the other hand, as per OpenAI, Terra can deliver the same performance as GPT-5.5 but at half the cost, while Luna is the most cost-efficient model.

The US government has stepped up scrutiny of advanced AI releases over concerns that the technology could be misused by military or intelligence actors in China, Russia or other countries of concern, including for cyberattacks.

Keep in mind that the US is not the only country that is evaluating limited access to its AI models for foreigners. As per reports, Chinese authorities are exploring such bans for advanced AI models developed by its AI labs too.

OpenAI’s plans to generally release GPT-5.6 comes shortly after the US government lifted curbs on Anthropic’s latest Fable and Mythos models. In June, foreigners, including Anthropic’s own employees, were banned from accessing the models over cybersecurity concerns.

The increased attention on advanced AI models came following an executive order signed by US President Donald Trump in June. The order set out a voluntary framework for AI developers to offer "covered frontier models" to the US government for up to 30 days before releasing them to trusted partners.

- Ends