OpenAI Limits Launch of New GPT-5.6 Models Following U.S. Government Request

New Delhi, June 27: OpenAI, a leading artificial intelligence (AI) company in the United States, has decided to restrict the launch of its new GPT-5.6 models for now. The company announced that the new AI models, named ‘Sol’, ‘Terra’, and ‘Luna’, will initially be released as a limited preview for select trusted partners. A broader rollout is planned in the coming weeks.
According to OpenAI, the new GPT-5.6 models feature significant improvements in reasoning, coding, and cybersecurity. The company also revealed that it informed the U.S. government about the capabilities of these models before the launch. Following the government’s request, the decision was made to limit the launch to establish a comprehensive framework for evaluating advanced AI models under the recently issued cybersecurity executive order.
OpenAI clarified that these models will be available on the ChatGPT, API, and Codex platforms in the next few weeks. However, the company expressed that it does not want this preview process to become a permanent rule imposed by the government before launches.
OpenAI has described ‘GPT-5.6 Sol’ as its flagship model for the most challenging scientific tasks. ‘Terra’ is presented as a balanced option for enterprise and developer tasks, offering capabilities similar to GPT-5.5 at nearly half the cost. ‘Luna’ is touted as the fastest and most affordable model, designed for low-cost applications.
The company stated that ‘Sol’ will be available on Cerebras hardware for select customers starting in July, operating at speeds of up to 750 tokens per second. It includes features like maximum logical effort and ultra mode, enabling it to handle complex coding, research, and multi-step tasks more effectively.
Cerebras Systems is a leading artificial intelligence hardware company.
Meanwhile, U.S. AI company Anthropic has also received approval to make its powerful Mythos ‘5 model’ available again to select trusted partners after addressing national security concerns. Just two weeks ago, the U.S. government had restricted access to the ‘Mythos 5’ and ‘Fable 5’ models for foreign nationals due to security reasons.


