OpenAI releases GPT-5.6 with three models: Sol at $5/$30 per 1M tokens, Terra, and Luna

OpenAI launched GPT-5.6 on June 26, 2026, introducing three models in limited preview: Sol (flagship), Terra (medium-tier for high-volume work), and Luna (fast and affordable). The company claims the models excel at coding, cybersecurity, and biology, with improved performance on long-horizon agentic AI tasks.

Pricing and model variants

GPT-5.6 Sol is priced at $5 input/$30 output per million tokens—nearly half the cost of Anthropic's Claude Fable 5 ($10/$50 per million tokens). Terra costs half of Sol's price, while Luna costs less than half of Terra's price. Specific pricing for Terra and Luna was not disclosed.

Sol includes two additional modes: "max" mode for deeper reasoning and "ultra" mode for leveraging sub-agents, similar to capabilities demonstrated by OpenClaw.

Security focus and government oversight

OpenAI dedicated most of its announcement to safety measures, stating that GPT-5.6 is "trained to refuse prohibited cyber assistance, including when users attempt to disguise their intent or jailbreak the model." According to OpenAI, Sol "is better at helping people find and fix vulnerabilities than reliably carrying out end-to-end attacks" and does not cross the cyber-critical threshold under OpenAI's preparedness framework—though the company revised this framework in April 2026, removing some previous study areas.

The company claims Sol has its "most robust safety stack to date" with "strengthened protections for higher-risk activity, sensitive cyber requests, and repeated misuse." OpenAI allocated approximately 700,000 A100e GPU hours to automated red-teaming and engaged third-party testers, who will continue testing for two weeks.

The release follows reports that the Trump administration requested involvement in the approval process. OpenAI stated that the administration will approve customers case-by-case during the preview period. The company noted that "safeguards may occasionally intervene on legitimate work, particularly in dual-use areas where defensive and offensive activity can initially look similar."

Availability timeline

OpenAI said the model suite will be generally available "in the coming weeks," emphasizing its belief in "broad access." The company stated it cooperated with the US government ahead of launch but expressed that such government access processes "should not become the long-term default" as they "keep the best tools from users, developers, enterprises, cyber defenders, and global partners who need them."

What this means

The aggressive pricing—Sol at nearly half of Claude Fable 5's cost—positions GPT-5.6 as a direct challenge to Anthropic's high-end offerings. The addition of sub-agent capabilities in "ultra" mode suggests OpenAI is pushing toward more autonomous AI systems, though government oversight during this release indicates heightened regulatory scrutiny around cybersecurity capabilities. The case-by-case approval process during preview represents an unprecedented level of government involvement in a commercial AI model release, raising questions about future deployment friction for frontier models with dual-use capabilities.