GPT-6 Luna
Language modelCard summarizedOur most efficient model for focused, high-volume tasks.
Active (latest)● Re-checked 12 Oct 2026 · 88 facts · 0 corrections · 1 unconfirmedSource: GPT-6 Astra System Card, Appendix A: GPT-6 Sol, GPT-6 Luna (175 pp)
Use it via API
Where to call it
- OpenAI API
gpt-6-luna
Supports
- Image input
- Streaming
- Structured outputs
- Reasoning control
- Tool calling
- Web search
- Prompt caching
- Batch API
Good for
- Focused, high-volume tasks at the lowest cost in the family
- Low-latency use with reasoning effort none
- Fewer factual errors than GPT-5.6 Luna, especially at low reasoning settings
Overview
GPT-6 Luna
Our most efficient model for focused, high-volume tasks.
- Cost
- $0.10 in
- Context
- 1.05M
- Input / output
- $0.10 / $0.50
- Max output
- 128K
- Knows up to
- May 2026
- Released
- 22 Sep 2026
OpenAI's fastest and most cost-efficient GPT-6 model, added to the GPT-6 Astra system card as an appendix. It is treated as High (not Critical) in cybersecurity and in biology and chemistry, and gets the same safeguards as GPT-5.6 Sol and Luna. Its safety scores improve on GPT-5.6, but some of that reflects a broader tendency to refuse. p.120, p.124
gpt-6-luna- Input / output per 1M
- $0.10 / $0.50
- Context window
- 1.05M
- Max output
- 128K
- Knows up to
- May 2026
- Reasoning
- Configurable (effort none → max)
- Weights
- Closed
About this card. A model card is the maker's own account of a model: what it's for, how it was trained and tested, and where it falls short. This page condenses OpenAI's GPT-6 Astra System Card, Appendix A: GPT-6 Sol, GPT-6 Luna. Every claim carries its source page.