GPT-5.6 Luna is the cost-efficient member of the GPT-5.6 trio, sitting beneath Sol and Terra as the lane tuned for high-volume, everyday work such as summarization, quick answers, and lightweight reasoning. OpenAI launched the family for general availability on July 9, 2026, after a limited preview, and Luna is positioned for situations where speed and price matter more than maximum reasoning depth. In ChatGPT it became the default model for Free users with unlimited text chats, with a Think button offered for harder questions that need extra reasoning, making it the everyday entry point into the family.

The model inherits the broader GPT-5.6 training focus on getting more useful work from every token, which carries through to Luna in the form of improved prompt caching behavior with explicit cache breakpoints and a 30-minute minimum cache lifetime that make caching more reliable for production agents. Independent reviewers describe Luna as the fastest, lowest-cost option in the family, well suited to coding-agent workloads that previously ran on older models, while the August update also exposed Luna through the API for high-volume integrations. For teams that need a balance of responsiveness, predictable caching, and low per-token cost for routine queries, Luna is the practical default inside the GPT-5.6 lineup.