gpt-5 6 — gpt-5 6 — OpenAI has publicly launched GPT-5.6, its most advanced AI model, following a delay last month prompted by US government requests over heightened national security concerns regarding the potential misuse of powerful AI technologies.
The release comes as the United States intensifies its efforts to develop cutting-edge AI models that experts warn could significantly accelerate sophisticated cyberattacks, particularly in sectors that depend on complex and often outdated technology systems.
In recent months, Washington has increased scrutiny over advanced AI model releases, aiming to identify and mitigate potential threats that could arise from the technology’s use by military or intelligence entities in other countries. This heightened vigilance is part of a broader strategy to ensure that the advancements in AI do not compromise national security.
In a notable move, OpenAI’s competitor, Anthropic, previously disabled its most advanced models, Mythos 5 and Fable 5, in response to a US government export control order issued on June 12. However, these restrictions were lifted last week after Anthropic introduced certain safeguards to address these concerns.
Axios reported that the Trump administration approved the broad launch of GPT-5.6 after additional testing and discussions between OpenAI and government officials. This development underscores the complex interplay between technological innovation and regulatory oversight.
Initially, OpenAI limited access to GPT-5.6, allowing only a select group of vetted partners to engage with the new model, which was disclosed to authorities. Alongside GPT-5.6 Sol, OpenAI also introduced two lower-cost models, Terra and Luna, as announced in an X post late on Tuesday.
Prior to the launch, OpenAI highlighted the improved capabilities of GPT-5.6 in areas such as coding, biology, and cybersecurity. In a preview shared in late June, the company indicated that GPT-5.6 Sol was competitive with Anthropic’s Mythos Preview on the ExploitBench cybersecurity benchmark, illustrating its advanced functionality.
The launch of these models comes at a time when the AI sector is rapidly evolving. On the same day as OpenAI’s announcement, billionaire Elon Musk revealed that his company, SpaceXAI, was also making its leading model, Grok 4.5, available to the public, further intensifying competition in the AI landscape.
Amid these developments, national security concerns remain at the forefront. President Trump signed an executive order establishing a voluntary framework for AI developers, allowing them to submit “covered frontier models” to the US government up to 30 days before releasing them to trusted partners. This order aims to ensure that cutting-edge technologies are carefully monitored for potential risks.
Despite the lifting of export controls for Anthropic’s Fable model, the Mythos model, which is tailored for cybersecurity professionals, continues to be restricted to select “trusted” US organisations. Anthropic has cautioned that achieving complete robustness against jailbreaks in AI models is “probably impossible,” highlighting ongoing challenges in the field.
