OpenAI Limits Its GPT-5.6 Sol, Terra, and Luna Models to Trusted Partners at the US Government's Request
OpenAI began a limited preview of its GPT-5.6 models — Sol, Terra, and Luna — restricting access to a small group of government-approved partners at the request of the US government.
Overview
OpenAI has begun a limited preview of GPT-5.6, a new three-model family, but is releasing it only to “a small group of trusted partners” following a request from the U.S. government, according to TechCrunch. The lineup is made up of Sol, Terra, and Luna, and the company has objected publicly to the access restrictions even as it complies with them.
What We Know
The GPT-5.6 family spans three tiers. Sol is the flagship and, in Engadget’s description, OpenAI’s “strongest model yet,” while Terra is a balanced option for everyday use and Luna is “the company’s lowest cost model,” according to Engadget. OpenAI says Terra offers similar performance to GPT-5.5 despite being twice as cheap, per Engadget.
Pricing scales with capability. Sol costs $5 per million input tokens and $30 per million output tokens, Terra is priced at $2.50 and $15, and Luna is the cheapest at $1 and $6, according to TechCrunch. OpenAI says Sol is slightly better at coding workflows than Anthropic’s Claude Mythos 5 while using a third of the output tokens, TechCrunch reported.
The unusual part of the launch is who can use the models. OpenAI is previewing GPT-5.6 only to partners “whose participation has been shared with the government,” according to TechCrunch. Federal leaders will be “approving access customer by customer during this preview period,” with a broader release expected “a couple of weeks later,” Engadget reported, citing a staff memo from OpenAI chief executive Sam Altman. OpenAI has said the models will become available more broadly in the coming weeks, per Engadget.
The staged rollout involves the White House Office of the National Cyber Director and the Office of Science and Technology Policy, along with Commerce Secretary Howard Lutnick, according to Engadget. The arrangement follows an executive order from President Trump requesting that AI companies voluntarily submit powerful models for federal review before public release, according to Engadget; that review can run up to 30 days before a launch, per TechCrunch.
OpenAI has been candid about its discomfort with the process. “We don’t believe this kind of government access process should become the long-term default,” the company wrote, according to TechCrunch. Altman told staff that the arrangement is not OpenAI’s preferred long-term model and that the company will work with the government and others in the industry on a more sustainable approach for future releases, per Engadget.
What We Don’t Know
OpenAI has not published a firm date for the general release, describing the timeline only as the coming weeks. The precise number of partner organizations admitted during the preview, and the criteria federal reviewers are using to approve each customer, remain undisclosed in the coverage reviewed here.
Analysis
The GPT-5.6 preview extends a pattern of the U.S. government gating access to the most capable AI models before they reach the public. The Trump administration previously ordered Anthropic to remove foreign-national access to its Claude Fable 5 model, according to TechCrunch; The Machine Herald previously reported that Washington later let Anthropic re-release its Mythos 5 model to trusted partners after an export block. With OpenAI now routing a flagship launch through a customer-by-customer federal approval process, the question is whether such reviews stay exceptional or harden into a standing gate on frontier releases — an outcome OpenAI itself says it wants to avoid.