OpenAI today announced a limited preview of its next-generation GPT-5.6 model series, introducing three distinct, capability-tiered models: Sol, Terra, and Luna. The rollout marks a strategic shift toward multi-agent architectures and deep-reasoning configurations, aimed at re-engineering developer and enterprise workflows.
Initially, approximately 20 trusted organizations—coordinated with the U.S. government—gain access to the new series. The flagship model, GPT-5.6 Sol, is priced at $5.00 per million input tokens and $30.00 per million output tokens, representing a major step-change in performance for long-horizon coding and cybersecurity tasks.
This launch reflects an unusual approach to AI deployment, with OpenAI aligning its release framework with the White House ahead of a broader public launch. Enterprise buyers will encounter real-time safety interventions, mandatory compliance parameters, and structured token caching systems in this new environment.
The core architectural evolution centers on how compute is allocated during inference, enabling the models to handle complex, multi-step reasoning more efficiently than prior generations. This could reshape competitive dynamics in the AI marketplace, as rivals like Anthropic and Google's DeepMind also push toward more agentic AI systems.
The limited preview suggests OpenAI is prioritizing safety and regulatory alignment over rapid market expansion. Industry observers will watch for early feedback from preview partners, as well as any signals from the U.S. government about potential regulatory frameworks for advanced AI models.