Faster and leaner in the family
OpenAI positions Luna as the fastest, most affordable GPT-5.6 option—suited to moving large volumes of short jobs off heavier models.
OpenAI · GPT-5.6 Luna · July 2026
GPT-5.6 Luna — lightweight model built for low latency and high-frequency calls
- Million-token context
- Extra-long single replies
- Fastest posture in the family
- Built for high-throughput loads
What is GPT-5.6 Luna?
GPT-5.6 Luna is OpenAI’s speed-oriented GPT-5.6 model, released in July 2026. It targets chat drafts, classification, summarization, and short-turn automation: within the family it prioritizes responsiveness and throughput, while still clearly stronger than earlier lightweight generations. Everyday balanced work belongs on Terra; the heaviest long-horizon jobs belong on Sol. On iMini Agent, pick GPT-5.6 Luna to use it.
- Vendor
- OpenAI
- Released
- July 2026
- Context
- 1M tokens
- Max output
- 128K tokens
- Deep reasoning
- Tunable · lightweight posture
- Best as
- High-turnover and short-turn primary
Where it shines
Speed, throughput, and the shared family surface—boundaries to know before putting Luna in a pipeline.
Still keeps million-token context
Context and output ceilings match the family, so short-turn jobs can still carry longer materials without jumping to Terra or Sol first.
Fits the front of a pipeline
Classification, summarization, drafts, and routing can run on Luna first; escalate to Terra or Sol when deeper checking is needed.
Shares the generation’s tool surface
Hooks into GPT-5.6 reasoning and tool-use capabilities so Agent flows can switch tiers step by step.
Official evaluations
Results OpenAI published with GPT-5.6. The shared x-axis is output-token volume. Luna is the family’s fastest tier, yet several scores still sit above the prior flagship.
Three common workflows
GPT-5.6 Luna fits short turns, high concurrency, and relatively controllable failure cost.
Drafts & rewrites
For first drafts and tone rewrites of email, explainers, and product copy. Ship a draft quickly, then proof critical spots on a heavier model.
Classification, extraction & summarization
For labels, field extraction, and meeting-note summaries. Focus on stable formats and high throughput.
Short-turn agent automation
For tool calls with few steps and a clear goal. Hard-code success criteria; send root-cause work and repo-scale changes to Sol.
How to choose vs GPT-5.6 Terra and GPT-5.6 Sol
Three tiers in one generation. The differences are depth, throughput, and the failure cost you can accept.
| Dimension | GPT-5.6 Luna | GPT-5.6 Terra | GPT-5.6 Sol |
|---|---|---|---|
| Context | 1M tokens | 1M tokens | 1M tokens |
| Reasoning | Lighter and faster | Balanced depth | Max depth · ultra available |
| Coding / agents | Short-turn drafts & light automation | Day-to-day coding & mid-weight agents | Heaviest long-horizon coding |
| Long tool runs | High turnover, shorter threads | More balanced completion and throughput | Flagship depth + ultra |
| Speed posture | Responsiveness and throughput first | Everyday primary favors balance | Completion quality first |
| Prefer when | Drafts, classification, and short-turn automation | Everyday coding and high-volume business | Highest failure-cost, heaviest jobs |
Using GPT-5.6 Luna on iMini
Free quota, model naming, key specs, and how it differs from Terra and Sol.
Use GPT-5.6 Luna free on iMini Agent
Million-token context, faster responses, built for high-turnover work.
Open GPT-5.6 LunaNo OpenAI API key required.