The U.S. Department of Commerce approved OpenAI on July 8 to broadly release its latest flagship model, the GPT-5.6 series; previously, GPT-5.6 Sol was only available for limited preview to trusted partners who had been informed to the government, and could only be accessed by customers approved by the Trump administration. OpenAI stated that Sol has not yet met its internal "cyber-critical" standard, but acknowledged that benchmarks cannot predict the possible uses of the model when combined with other tools.
GPT-5.6 Series Three Models: Sol's Functional Positioning and "Cyber-Critical" Standard Assessment
According to OpenAI's public statement, the GPT-5.6 series includes the following three models:
GPT-5.6 Sol: Positioned as the most powerful model in the series, with improvements in coding, biology, and cybersecurity tasks; OpenAI stated that Sol is better at helping people find and fix software vulnerabilities rather than reliably executing full cyber attacks
GPT-5.6 Terra: A member of the GPT-5.6 series
GPT-5.6 Luna: A member of the GPT-5.6 series
OpenAI stated that GPT-5.6 will initially be open to trusted partners via API and Codex, and as of early July has not been launched in the ChatGPT preview version, with no official release date announced. Additionally, OpenAI acknowledged that benchmarks cannot predict all possible uses of the model when combined with other tools.
Trump Administration June Executive Order Framework: Legal Distinction Between Voluntary Framework and Mandatory Approval
According to a White House fact sheet, the executive order issued by the Trump administration in June established a "voluntary framework" for regulated frontier AI models, allowing the government to provide secure early access for trusted partners; the White House also clearly stated that the executive order "does not authorize mandatory licensing, pre-approval, or approval of the release of AI models."
Critics have expressed concerns that selective early access could evolve into a mechanism for the government to quietly control who gets early access to the latest AI tools; the above are the views of critics, and the official government position is based on the text of the executive order.
Comparison of Anthropic Fable 5 and Mythos 5 Cases
The cases of Anthropic's Fable 5 and Mythos 5 models serve as a useful contrast for GPT-5.6 access management: On June 12, 2026, Anthropic stated that the U.S. government instructed it to suspend foreign citizen access to Fable 5 and Mythos 5; Anthropic said the order forced it to disable these models more broadly to comply.
Subsequently, a breakthrough emerged: the U.S. Department of Commerce canceled the corresponding export controls; Anthropic stated it would restore access, and Fable 5 reportedly came back online on July 1; Anthropic agreed to strengthen security measures, cooperate with the U.S. government in developing model release protocols, and report malicious activities related to these models; Reuters reported that the restrictions were lifted after Anthropic met with U.S. officials.
Frequently Asked Questions
When will GPT-5.6 be available on ChatGPT after the U.S. Department of Commerce approves its broad release?
According to OpenAI's statement, as of early July, GPT-5.6 has not been launched in the ChatGPT preview version, and the official release date has not been announced; GPT-5.6 is currently mainly accessible via API and Codex to some trusted partners. The specific launch timeline is subject to OpenAI's official announcements.
Does the Trump administration's intervention in AI model releases amount to mandatory approval?
According to the White House fact sheet, the June executive order clearly states that it "does not authorize mandatory licensing, pre-approval, or approval of the release of AI models," positioning it as a "voluntary framework"; critics, however, worry that selective early access could create a de facto access control mechanism. The above are the positions of various parties; the official text of the executive order and subsequent government statements shall prevail.
How did Anthropic reach an agreement with the U.S. government to restore access?
According to reports, Anthropic agreed to strengthen security measures, cooperate with the U.S. government in developing model release protocols, and report malicious activities related to Fable 5 and Mythos 5; Reuters reported that the restrictions were lifted after Anthropic met with U.S. officials. Specific terms of the agreement are subject to official statements from Anthropic and the U.S. Department of Commerce.
