GPT-6 Luna is OpenAI's most efficient GPT-6 model, built for focused, high volume tasks such as classification, extraction, routine drafting and simple coding. OpenAI released it in the API on September 22, 2026, alongside GPT-6 Sol, and it also powers GPT-6 in ChatGPT's Chat tab for Free and Go users.
Key facts
| Item | GPT-6 Luna |
|---|---|
| Maker | OpenAI |
| Released | September 22, 2026 |
| API model ID | gpt-6-luna |
| Context window | 1,050,000 tokens (up to 922,000 input) |
| Max output | 128,000 tokens |
| Knowledge cutoff | May 18, 2026 |
| Input / output | Text and image in, text out |
| API price per million tokens | $0.10 input, $0.01 cached input, $0.50 output |
Sources: OpenAI's GPT-6 Luna model page, the API changelog and the GPT-6 Sol and Luna announcement, checked October 24, 2026.
Pricing details
Luna costs $0.10 per million input tokens and $0.50 per million output tokens, a hundredth of GPT-6 Astra's rate. Cached input costs $0.01 and cache writes $0.125. Prompts over 272,000 input tokens cost twice the input and cache rates and 1.5 times the output rate. Batch and Flex cost half, Fast mode costs double, and EU data residency is available with Standard, Fast, Flex and Batch processing.
Where to use GPT-6 Luna
- API: on the OpenAI API. It is the only model for OpenAI's Decisions API, which became generally available on October 22, 2026. Reasoning effort runs from
nonetomax. - ChatGPT: it powers GPT-6 in the Chat tab for Free and Go users from October 8, 2026, appears in the desktop app for those plans, and is in Work and Codex for paid plans of ChatGPT.
How it compares
Its closest rival is Anthropic's Claude Haiku 5.5, which starts at the same $0.10 and $0.50 per million tokens. Our Claude Haiku 5.5 vs GPT-6 Luna guide compares them in depth, and GPT-6 Sol vs Astra vs Luna explains where Luna fits in OpenAI's lineup.
Who should use it
Use Luna for bulk and real time work where cost and speed matter more than depth. For longer agentic tasks, step up to GPT-6.1 Sol. See the release timeline for every GPT-6 launch.