GPT-6 Luna: Pricing, Context Window and Model Guide

What GPT-6 Luna is

OpenAI released GPT-6 Luna on September 22, 2026, together with GPT-6 Sol. OpenAI calls Luna its most efficient model for focused, high-volume tasks, and says it was trained with methods similar to GPT-6 Astra.

Pricing

Standard rates per million tokens are $0.10 for input, $0.01 for cached input, $0.125 for cache writes and $0.50 for output. Batch and Flex halve those again, and Fast mode doubles them. That is a twentieth of GPT-6 Sol's price and a hundredth of GPT-6 Astra's. OpenAI positioned the price as 50% below GPT-5.6 Luna's promotional pricing. As with the rest of the family, prompts over 272K input tokens are billed at twice the input and cache rates and one and a half times output for the whole request.

Context, output and reasoning

Luna has the same 1,050,000-token context window and 128,000-token maximum output as Sol. Its knowledge cutoff is May 18, 2026. Reasoning effort runs from none through low, medium, high and xhigh to max, with medium the default. Input is text and images, output is text.

Tools and where it runs

Luna supports the same tool set as the larger models: web search, file search, image generation, the code interpreter, a hosted shell, apply_patch, skills, computer use, MCP servers and tool search, on Chat Completions and Responses. It is available in the API and in ChatGPT Work and Codex, and OpenAI says Free and Go users can reach it in the desktop app.

How it measures up

On LMArena's text board GPT-6 Luna has 1441 Elo (rank 99, max effort); in coding, 1501 Elo in Coding (rank 80, max effort). That is twelve points behind GPT-6 Sol, at a twentieth of its price. Current figures with their confidence ranges are on our benchmarks page and update weekly.

Official Resources