OpenAI's GPT-5.6 Series Set for Global Rollout Following US Nod

OpenAI, the company behind the widely used ChatGPT, has confirmed that its most recent and powerful artificial intelligence model series will be made available to the public this Thursday. This announcement follows reports indicating that the United States government has given its approval for a more extensive launch.

Concerns have been raised regarding OpenAI's new GPT-5.6 offerings, alongside other advanced AI models like Anthropic’s Mythos series, particularly their alleged unprecedented capability to pinpoint vulnerabilities in code that could be exploited by hackers. Such capabilities have sparked national security anxieties. In response, OpenAI disclosed in late June that it had granted a select group of trusted US-based collaborators preliminary access to GPT-5.6, at the request of Washington.

Large language models (LLMs) form the core technology powering chatbots and various other AI tools, thanks to their immense capacity to process vast amounts of digital data. The GPT-5.6 series is structured into three distinct tiers: Sol, which is the company’s new flagship model; Terra, a mid-range variant designed for everyday tasks; and Luna, a rapid and cost-effective option.

On Tuesday, OpenAI posted on X, stating, "GPT-5.6 Sol, along with Terra and Luna, will launch publicly this Thursday. We’re expanding preview access globally now," without providing additional specifics. According to US news outlet Axios, which cited an informed source, the Trump administration reportedly granted the company the go-ahead for a broad release of GPT-5.6. This decision came after a period of technical evaluations and discussions between the company and government officials. AFP has reached out to OpenAI, the White House, and the US Department of Commerce for comments on the Axios report.

This development mirrors a similar situation involving OpenAI’s primary competitor, Anthropic, the startup responsible for the Claude chatbot. Just last week, Anthropic announced it would recommence access to its most potent AI models, Fable 5 and Mythos 5, after Washington lifted previous restrictions on their distribution. Prior to the emergence of Mythos, the administration of President Donald Trump had advocated for fewer regulations on AI companies, believing it would give the US an edge in the global AI race against China.

The government is currently in the process of defining criteria for AI models that will fall under new security regulations, in line with an executive order issued by the White House. OpenAI expressed its reservations in June, stating, "we don’t believe this kind of government access process should become the long-term default" because it "keeps the best tools" away from users, businesses, and other entities that could benefit from them. The company further noted that it is collaborating with Washington "to develop the cyber Executive Order framework and a repeatable process for future model releases."

OpenAI has indicated that once widely available, Terra will be priced at half the cost of its predecessor, GPT-5.5. This strategic pricing aims to secure a larger customer base amidst intense competition from rivals such as Anthropic and Google. Both OpenAI and Anthropic have confidentially submitted initial public offering (IPO) documents to US regulators, with aspirations for public listings valued at nearly $1 trillion, thereby intensifying the commercial stakes in the ongoing AI arms race between them.