OpenAI limits GPT-5.6 rollout after US government request, warns against permanent restrictions

OpenAI announced it will limit the initial release of its next-generation GPT-5.6 model family—Sol, Terra, and Luna—to a small group of trusted partners after a request from the Trump administration. The company made clear it disagrees with the move, stating that such government access processes should not become the long-term default. Sol, the most capable model, features enhanced agentic abilities in coding, biology, and cybersecurity, with a new “max” reasoning effort mode and an “ultra” mode that coordinates subagents for highly complex tasks. OpenAI claims Sol is slightly better than Anthropic’s recently banned Claude Mythos 5 at coding workflows while using only a third of the output tokens.

To address safety concerns, OpenAI says Sol includes its most robust security stack yet, with defenses against adversarial attacks and a design that favors defensive cybersecurity work over offensive exploits. The safety guardrails are embedded directly into the model’s core behavior rather than relying on a separate filter—a lesson from Anthropic’s Fable 5, which faced backlash for its over-cautious routing system. The three models are tier-priced: Sol at $5/M input and $30/M output tokens, Terra at half that, and Luna at $1/M input and $6/M output. OpenAI expects broader availability via ChatGPT, Codex, and the API in the coming weeks, pending a new executive order framework on cybersecurity.