OpenAI GPT-5.6 launches with Trump admin pre-approval; three pricing tiers (Sol/Terra/Luna)

OpenAI publicly launched all three GPT-5.6 variants (Sol, Terra, Luna) on July 9 after receiving Trump administration approval for wider release. Initially deployed to 'trusted partners' in late June per Trump's June AI cybersecurity order (which requires 30-day pre-release government review), OpenAI compressed the timeline after the Department of Commerce's Center for AI Standards and Innovation completed additional testing and met with company technical experts in DC. The pre-approval process sets a precedent for frontier model launches in 2026.

Sol is OpenAI's strongest model yet at $5/$30 per million tokens (input/output). Terra targets everyday use at $2.50/$15, claiming GPT-5.5 performance at 2x lower cost. Luna is the lowest-cost variant at $1/$6. The three-tier strategy mirrors Anthropic's Fable 5/Sonnet 5 stacking and suggests a market shift toward price-based segmentation by task. Terra's cost directly competes with Claude Sonnet 5 ($2/$10 introductory through August 31).

For practitioners, the government pre-approval process itself matters more than the model delta. If this becomes standard, frontier model deployment cycles will be constrained by regulatory review timelines rather than engineering readiness. Both OpenAI and Anthropic have filed confidentially for IPO in 2026; government access requirements could shape disclosure obligations and capital market valuations if investors demand regulatory risk quantification.