The Bottom Line: OpenAI is rolling out GPT-5.6 with three variants, with Terra delivering GPT-5.5 performance at half the cost.
OpenAI has announced a limited preview of three new models in the GPT-5.6 series: Sol as the flagship, Terra as a balanced general-purpose model, and Luna as a cost-effective variant. The models will be made generally available over the coming weeks.
OpenAI has launched a limited preview of the GPT-5.6 model family with three variants. Sol serves as the flagship model with the highest performance. Terra is a balanced model for everyday tasks with competitive performance relative to GPT-5.5, but costs 50 percent less. Luna delivers strong capabilities at the lowest price point in the portfolio.
The initial phase operates as a limited preview for a small group of trusted partners, whose participation has been notified to the U.S. government. OpenAI plans to make the models generally available over the coming weeks, underscoring its confidence in a broad access model.
For CTOs, the cost structure is particularly relevant: Sol costs 5 USD per 1 million input tokens and 30 USD per 1 million output tokens. Terra is priced at 2.50 USD input and 15 USD output. Luna is the most cost-effective offering at 1 USD input and 6 USD output. This tiered pricing makes it possible to evaluate different use cases in a differentiated manner — from high-performance production operations to budget-optimized automation.
Additionally, GPT-5.6 introduces improved prompt caching. Cache writes are charged at 1.25 times the uncached input rate, while cache reads continue to receive a 90-percent discount on the input rate. Caching includes explicit cache breakpoints and a minimum lifetime of 30 minutes, reducing token costs for repeated requests.
Source: simonwillison.net · Published June 26, 2026
Lumi AI News — AI-assisted curation pursuant to Art. 50 EU AI Act. Paraphrase and classification by Lumi News Pipeline v1.7.1.