# GPT-6 Luna

> GPT-6 Luna is OpenAI's most efficient GPT-6 model, for focused, high-volume tasks such as summarization, extraction and focused coding. It started rolling out on September 22, 2026 and costs $0.10 / $0.50 per 1M input/output tokens on the API, with a 1.05M-token context window.

URL: https://aistack.sh/model/gpt-6-luna
Vendor: OpenAI (product page: https://aistack.sh/tool/chatgpt)
Last checked against sources: 2026-09-24

## Specs

- Released: Sep 22, 2026
- Status: Rolling out
- API model ID: gpt-6-luna
- API price (in / out): $0.10 / $0.50 per 1M tokens
- Cache read: $0.01 per 1M tokens
- Context window: 1.05M tokens
- Plans: Plus, Pro, Business, Enterprise, Edu, Free and Go (desktop app)
- Platforms: OpenAI API

## Where you can use it

- ChatGPT (https://aistack.sh/tool/chatgpt): In ChatGPT Work on the web and the desktop app; Free and Go users get it in the desktop app.
- Codex (https://aistack.sh/tool/codex): In the Codex CLI, IDE extension and Codex cloud. The rate-limit switch prompt recommends it.
- API prices shown are OpenAI's short-context rates; long-context requests are billed higher.

## FAQ

### How much does GPT-6 Luna cost on the API?

$0.10 per 1M input tokens, $0.01 per 1M cached input tokens and $0.50 per 1M output tokens at OpenAI's short-context rates.

### Can free ChatGPT users use GPT-6 Luna?

Yes, in the desktop app. OpenAI's changelog says Free and Go users can access Luna there; Plus, Pro, Business, Enterprise and Edu get it in Codex and ChatGPT Work.

### What is GPT-6 Luna for?

OpenAI describes it as its most efficient model for focused, high-volume tasks, including summarization, extraction and focused coding.

## Sources

- [OpenAI: ChatGPT Work and Codex changelog](https://learn.chatgpt.com/docs/changelog), read 2026-09-24
- [OpenAI: ChatGPT and Codex models](https://learn.chatgpt.com/docs/models), read 2026-09-24
- [OpenAI API models](https://developers.openai.com/api/docs/models), read 2026-09-24
- [OpenAI API pricing](https://developers.openai.com/api/docs/pricing), read 2026-09-24