OpenAI has restricted access to its GPT-5.6 models to a small group of trusted partners following a request from the US government.
The rollout includes Sol, the flagship model; Terra, a balanced option for daily tasks; and Luna, a faster, lower-cost alternative. Despite Sol being the most powerful, the Trump administration has blocked access for all three. OpenAI stated the preview is limited to partners whose involvement has been shared with the government.
This move follows increased pressure on AI firms to restrict advanced systems. After Anthropic released its public model Fable 5, the administration ordered the company to remove access for any foreign national, forcing Anthropic to take the model down entirely.
Dean Ball, a former White House AI adviser and future OpenAI employee, argues the administration’s recent executive order has created an involuntary licensing regime for frontier AI. The order asks certain companies to voluntarily submit their most advanced models for review up to 30 days before release. Ball says this leads to heavy-handed restrictions because the government lacks clearly defined safety standards, which could cause endless launch delays.
These delays might give China an advantage in the AI race while jeopardising billions of dollars invested in AI infrastructure buildouts. OpenAI complied with the administration’s request this time but made it clear it was unhappy with the arrangement.
“We don’t believe this kind of government access process should become the long-term default,” reads a Friday blog post. “It keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them.”
OpenAI described the preview as a short-term step that will put GPT-5.6 on the path to broader availability in the coming weeks. The company is working with the administration to develop a new executive order framework on cybersecurity and a repeatable process for future model releases.
GPT-5.6 Sol specs
OpenAI says GPT-5.6 Sol is its strongest model yet, with improved agentic capabilities in coding, biology, and cybersecurity. Sol introduces a max reasoning effort mode and an ultra mode that uses coordinated subagents to solve highly complex tasks. This approach uses token usage heavily.
GPT-5.6 excels at several benchmarks, according to OpenAI. It performs slightly better at coding workflows than Anthropic’s Claude Mythos 5, which the Trump administration also effectively banned this month. OpenAI says GPT-5.6 Sol is competitive with Mythos preview but uses a third of the output tokens.
To address concerns about safety, OpenAI says Sol includes its most robust security stack yet. It is heavily hardened against adversarial attacks and intentionally optimised to favour defensive cybersecurity work over offensive exploits. In other words, it is designed to be hard to jailbreak, while prioritising showing users how to defend against exploits rather than how to hack into systems.
OpenAI also says its safety guardrails are built directly into the core model’s behaviour, rather than relying on a separate filter on top of it. The firm is likely trying to avoid the trap that caught Anthropic with Fable 5. When Fable 5 was available, its classifiers would route requests to an older model whenever they detected a high-risk topic like cybersecurity, biology, or chemistry. This over-cautious flow led to many false positives and user backlash.
While the GPT-5.6 models are initially available only to a select group of partners, OpenAI plans to make them more broadly available to people using ChatGPT, Codex, and the API soon.
GPT-5.6 comes in three sizes with tiered pricing. Sol costs $5 per million input tokens and $30 per million output tokens. Terra costs half that. Luna costs $1 and $6, respectively. OpenAI says it has also improved prompt caching to make repeated prompts cheaper and more predictable.
What it means
Developers and businesses face a new reality where launching advanced tools requires prior government approval. This creates uncertainty for companies planning infrastructure investments and delays access for users who need the latest capabilities immediately. OpenAI’s stance suggests this friction is temporary, but the precedent establishes a new barrier between innovation and public availability.
