OpenAI announces three next-generation frontier models — GPT-5.6 Sol + Terra + Luna in limited preview in Codex and API, Sol is most capable cybersecurity model yet with most robust safety stack yet
OpenAI announced three next-generation frontier models — GPT-5.6 Sol, Terra, and Luna — available as limited preview in Codex and the API for a small group of partners. GPT-5.6 Sol is OpenAI's most capable cybersecurity model yet, launching with OpenAI's most robust safety stack yet. The three-model trio operationalizes the government-gated paradigm at full Sol/Terra/Luna scope per US government request.
The substantive piece is the three-model trio framing at frontier-tier capability. Yesterday's coverage identified GPT-5.6 Sol as the limited-preview release with government-gating; today's expanded details name the full Sol + Terra + Luna trio + Sol's cybersecurity-focused positioning + most-robust-safety-stack-yet framing. The trio scope establishes that OpenAI is releasing three differentiated frontier capabilities simultaneously rather than a single model.
The competitive read against yesterday's government-gated AI paradigm HC is that the H2 2026 frontier-AI release pattern now operationalizes at full-product-portfolio scope. Government-gated isn't just for one flagship model; it applies to differentiated frontier capability portfolios. The H2 2026 to 2027 procurement landscape for frontier AI operates entirely within government-approved-partner architecture for OpenAI + Anthropic offerings.
OpenAI — Introducing GPT-5.6 Sol cybersecurity-focused frontier model → · Thurrott — OpenAI Launches Next-Gen GPT 5.6 Models in Limited Preview →