GPT-5.6 Luna

GPT-5.6 Luna is a cost-efficient model from OpenAI for fast, high-volume workloads. It is the first version of the gpt-luna line, released on 2026-07-09. It has a context window of 1,050,000 tokens and a maximum output of 128,000 tokens. Pricing starts at $0.06 per million input tokens and $0.37 per million output tokens, with cached input at $0.02 per million tokens. The weights are not open. It supports reasoning mode, tool calling, and accepts image, PDF, and text inputs, returning text. Knowledge cutoff is 2026-02-16.

First version of its line in the registry.

Range across the hosts that serve it: up to $1.00 per 1M input tokens.

• ·1.05M token context window
• ·128K token max output
• ·Priced from $0.06 per M input tokens
• ·Supports reasoning and tool calling
• ·Accepts image, PDF, and text input

• Reach for it for high-volume text generation workloads
• Reach for it for processing long documents with 1M+ token context
• Reach for it for multimodal inputs like images and PDFs
• Reach for it for cost-efficient batch inference

32 hosts serve this model at the price above · the maker's documentation

• 8Arena Review: GPT-6 Luna (Max) Ranks 23rd on Agent Arena with a Single Task Cost of Only $0.052026-09-29
• 8Claude Opus 5.5, GPT-6 Sol, GPT-6 Luna, and a new price war2026-09-23

2 signals of the archive mention this model, each scored against the published bar. This list is rebuilt from the archive at every build, so it cannot go stale.

The registry keeps the six most recent versions of a line. Older ones stay in the file and can be surfaced without a rebuild.

What is the context window?

The context window is 1,050,000 tokens.

What is the pricing?

Input from $0.06 per million tokens, output from $0.37 per million, cached input from $0.02 per million.

Does it support image input?

Yes, it accepts image, PDF, and text inputs, and returns text.