GPT-5.6: OpenAI Held Back by Washington Before Public Launch

GPT-5.6: A Limited Release Due to U.S. Regulations

OpenAI recently unveiled its latest series of artificial intelligence models, named GPT-5.6. Among these models, GPT-5.6 Sol stands out as the most powerful ever designed by the company. However, contrary to what one might expect from such a technological advancement, these models are not yet accessible to the general public on ChatGPT, at the request of Washington. Indeed, a directive from the U.S. government has imposed a limited pre-launch to a select group of partners. This decision comes as Claude Fable 5, another AI model, also remains inaccessible for similar reasons.

The date of June 26 marked the official presentation of GPT-5.6 by OpenAI. Despite the excitement generated by this announcement, access to these models is currently reserved for a limited number of carefully selected partners. This restriction is the result of ongoing collaboration between OpenAI and the U.S. government. OpenAI justified this approach by stating, “As part of our ongoing collaboration with the U.S. government, we presented our plans and the capabilities of our models prior to today’s launch. At their request, we are starting with a pre-launch phase reserved for a small group of trusted partners, whose participation has been communicated to the government, before proceeding to a broader deployment.”

A Public Deployment in the Near Future

OpenAI has announced that the general public could access the GPT-5.6 models in a few weeks. Sam Altman, the CEO of OpenAI, expressed on X his agreement with the necessity of a gradual deployment of new AIs, especially as they become more sophisticated. However, he also emphasized the importance of developing a more efficient process for the future.

The new AI, Sol, is not only smarter and more efficient, but it is also offered at the same price as its predecessor, GPT-5.5. Despite this advancement, the initial launch is limited, in accordance with the request of the U.S. government. Altman expressed concerns about the sustainability of this approach, highlighting that it could deprive many users, developers, businesses, cybersecurity professionals, and international partners of the necessary tools for their activities.

Towards a More Flexible Regulatory Framework

OpenAI is working closely with the U.S. government to establish a framework that would facilitate the launch of its future models. The company hopes that this cooperation will help overcome current obstacles and prevent such restrictions from becoming the norm. In a statement on OpenAI's blog, it was mentioned that this type of access procedure could limit innovation and access to cutting-edge technologies for developers and businesses.

Powerful Yet Secure Models

Despite the current restrictions, OpenAI remains optimistic about the imminent lifting of limitations on GPT-5.6 Sol. The company has implemented rigorous security measures to prevent any malicious use of its models. Although the new AIs bring significant improvements in cybersecurity, they are not considered to pose a high risk according to OpenAI's internal assessments.

In addition to GPT-5.6 Sol, OpenAI has also developed other models such as GPT-5.6 Terra and GPT-5.6 Luna. Terra is designed to offer an optimal balance between performance and cost, competing with GPT-5.5 while being half the price. Luna, on the other hand, is a fast and affordable model aimed at meeting the needs of a broader audience.