US-based AI company OpenAI is set to introduce new GPT-5.6 models named Sol, Terra, and Luna. These models will initially have a limited preview for trusted partners, with a broader release planned in the near future. OpenAI’s latest models promise advancements in reasoning, coding, and cybersecurity, as stated by the company.
OpenAI informed the US government about the capabilities of GPT-5.6 before the launch. The company agreed to a limited release at the government’s request while federal authorities work on a comprehensive framework for evaluating advanced models under a recent cybersecurity executive order.
The upcoming models are expected to be accessible across ChatGPT, the API, and Codex in the upcoming weeks. OpenAI clarified that it does not want government preview requirements to become a permanent standard for future AI releases. The launch of these models follows heightened scrutiny of advanced AI models in the United States.
OpenAI has detailed the characteristics of the new models. Sol is positioned as the flagship model for demanding scientific workloads, while Terra offers a balanced option for enterprise and developer tasks at a reduced cost compared to its predecessor, GPT-5.5. Luna, on the other hand, is introduced as the fastest and most cost-effective model suitable for budget-sensitive applications without compromising core capabilities.
Sol, the flagship model, will be accessible on Cerebras hardware starting July for selected customers. It boasts impressive inference speeds of up to 750 tokens per second. Sol features Max Reasoning Effort, an enhanced capability allowing more time for complex problem-solving. Additionally, its Ultra Mode utilizes multiple specialized sub-agents to enhance performance in coding, research, and multi-step tasks.
In a related development, US-based AI company Anthropic has received approval from US authorities to reinstate limited access to its powerful Mythos 5 AI model for specific trusted partners. This decision follows the resolution of national security concerns raised by the government. Earlier, Anthropic faced restrictions after concerns were raised that security protocols could be bypassed, prompting a temporary ban on providing access to foreign nationals.
