GPT-5.6 Luna is positioned as the fastest and most cost-efficient tier within the GPT-5.6 family, sitting beneath the higher-capability Sol tier and the mid-range Terra tier. OpenAI designed Luna specifically for high-volume workloads where token economics matter as much as raw capability, making it well suited to automated agentic pipelines that iterate on planning, tool calls, and revisions. Its release coincided with an upgrade to Auto-review in the ChatGPT app and Codex CLI, extending this cost-optimized tier into automated workflows that previously required manual approval steps for lower-risk work. The model's broad availability across ChatGPT, Codex, and the OpenAI API means teams can route routine or repetitive tasks to Luna while reserving richer tiers for more demanding problems.
Luna supports multimodal input alongside text generation, and includes reasoning, tool calling, and structured output capabilities, which together make it practical for coding agents and other tool-using assistants that need to interpret attachments and act on results. A tiered pricing reduction formalized around the model's launch lowered Luna's per-token cost substantially compared with earlier positioning, reinforcing its role as the entry-level option for sustained, high-throughput usage. This combination of affordability, multimodal comprehension, and tool-aware reasoning lets organizations shift large volumes of routine agentic work to Luna without sacrificing the functional surface area needed for vision inputs, function calls, or structured responses.
