
OpenAI is staggering the launch of its flagship GPT-5.6 Sol model after a request from the Trump administration, requiring customers to obtain U.S. government approval before access. The move reflects escalating government concern about cyber capabilities in advanced AI systems, following similar export controls imposed on Anthropic in early June. OpenAI describes the step as voluntary and temporary while the administration develops a formal framework for vetting models before release.
Summaries like this, in your inbox every morning.
Sign up free →What happened
OpenAI is staggering the rollout of its newest model, GPT-5.6 Sol, after a request from the Trump administration. Customers must first be cleared by the U.S. government to access it. The company says Sol is its strongest model yet, able to complete 50% of long-running professional tasks and tops all previous OpenAI models on coding capabilities. Initial users are customers approved by the government, with the list expanding next week.
Why it matters
The move reflects growing U.S. government concern about advanced cyber capabilities in frontier AI models. Advanced cyber capabilities displayed by Anthropic's Mythos and OpenAI's GPT-cyber have caused concern in Washington. This is the second time in a month a frontier lab's best model has been held back over capability concerns—in early June, the Commerce Department issued export controls on Anthropic that forced the lab to cut off foreign access to two of its top models. The staggered approach may signal how the government will oversee powerful AI releases going forward.
What to watch
Sol is priced at $5 per million input tokens and $30 per million output tokens, compared to Terra at $2.50 and $15, respectively, and Luna at $1 and $6. OpenAI says it spent 700,000 GPU hours testing for vulnerabilities and humans will conduct two more weeks of tests before broader launch. The company emphasized that Sol did not cross the 'Cyber Critical' threshold in its 'Preparedness Framework'—in tests with Firefox and Chrome, it found the seeds of an exploit but did not produce a working one.
No discussion yet for this article
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
1 minute a day. The AI essentials.
200+ sources · Email / LINE / Slack