GPT-6 Luna
Positioning
GPT-6 Luna is an efficient reasoning model for focused, high-volume tasks. OpenAI also describes the version used in ChatGPT as tuned for everyday conversation. (developers.openai.com)
Capabilities and uses
- The API accepts text and image inputs and produces text. It supports adjustable reasoning effort, streaming, function calling and structured outputs. (developers.openai.com)
- Suitable uses include everyday questions, coding and business workflows. In ChatGPT, GPT-6’s Intelligent UI can present answers using graphics, charts, forms and interactive elements; examples include explanations, comparisons and simple tools. This is a ChatGPT experience, not a guarantee that an API response will render an interface. (openai.com)
- The Responses API supports built-in tools and function calling. In Chat Completions, function calling requires reasoning effort to be set to
none. (developers.openai.com)
Limits
The documented API context window is 1,050,000 tokens, with a maximum output of 128,000 tokens and a knowledge cutoff of May 18, 2026. Audio and video inputs are not supported by this model; fine-tuning is not supported. OpenAI says Intelligent UI still needs improvements in design judgment and in the range of interfaces it can create. (developers.openai.com)
Release and availability
The supplied announcement was published on October 7, 2026. It said the ChatGPT Chat rollout for Free and Go users, powered by Luna, would begin on October 8, 2026. This was not the model’s initial API release: OpenAI’s API changelog records Luna’s release on September 22, 2026. (openai.com)
API pricing
OpenAI direct API prices are in USD per 1 million tokens, checked November 16, 2026. These are API charges, not ChatGPT subscription prices. (developers.openai.com)
| Standard processing | Input | Cached input | Output |
|---|---|---|---|
| Up to 272K input tokens | $0.10 | $0.01 | $0.50 |
| More than 272K input tokens | $0.20 | $0.02 | $0.75 |
For prompts over 272K input tokens, the higher rates apply to the full request. Batch and Flex cost 50% of the applicable Standard rates; Fast costs twice the applicable rates. Regional processing adds a 10% premium where available. Cache writes and tools may incur additional charges; consult the official pricing pages for conditions. (developers.openai.com)