OpenAI announced that it is beginning a limited preview of GPT-5.6 today, with the Trump administration limiting the public launch. OpenAI has agreed to hold back broad release while the government conducts review, making the models available only to a small group of trusted partners for now. General availability for Sol, Terra, and Luna is planned in the coming weeks once further testing is complete.
GPT-5.6 – advanced reasoning and pricing
GPT-5.6 Sol is OpenAI’s strongest model to date, with improvements in coding, biology, and cybersecurity. The model introduces new “max” and “ultra” reasoning effort options; ultra mode uses subagents to accelerate complex work, allowing Sol to handle problems that would require manual coordination.
Pricing reflects OpenAI’s competitive positioning across all three tiers:
- Sol: $5 input / $30 output per 1M tokens
- Terra: $2.50 input / $15 output per 1M tokens
- Luna: $1 input / $6 output per 1M tokens
Terra performs similarly to GPT-5.5 but is 2x cheaper, making it OpenAI’s primary option for general users and developers not requiring Sol’s advanced capabilities. ChatGPT’s competitive positioning has long hinged on balancing capability with cost-efficiency, and Luna’s pricing positions it as a direct competitor to lower-cost alternatives in the market.
Safety and government approval
OpenAI says GPT-5.6 Sol has the company’s “most robust safety stack to date,” with protections for high-risk activity, sensitive cyber requests, and misuse. The company has built explicit guardrails targeting cybersecurity research and biological work, areas where advanced reasoning could enable harmful applications.
In its announcement, OpenAI pushed back against the administration’s request to hold the model back, arguing that government AI access should not become the long-term default. The tension illustrates a core challenge for AI companies: demonstrating safety rigor while resisting indefinite delays that could cede market advantage to international competitors.
Prompt caching and infrastructure
OpenAI’s latest platform updates include predictable prompt caching with explicit cache breakpoints and a 30-minute minimum cache life. Cache writes are billed at 1.25x the uncached input rate, letting developers reduce costs on repeated contexts while maintaining clarity around billing.
