OpenAI GPT-6 Luna

GPT-6 Luna is OpenAI's budget-tier entry in the GPT-6 family, released alongside the mid-tier Sol and positioned beneath the flagship Astra. According to OpenAI, Luna was trained using methods similar to those used for Astra, with the stated aim of bringing improvements in professional work, factual accuracy, coding, computer use, and alignment to a faster and more affordable model. Luna is aimed at high-volume office tasks with clear objectives, such as summarizing documents, extracting information, and answering quick questions, while still being capable of meaningful coding work. OpenAI calls it the cheapest model that can still do real coding.

In practice Luna is tuned for cost-efficient automation and subagent work. At max effort it reaches 66.6% on the DeepSWE long software-engineering benchmark for an estimated $0.22 per task, matches Sol on key coding tests, and stays under $0.30 per task across OpenAI's benchmark charts. It supports six reasoning effort levels and defaults to medium, with the Responses API recommended for agent loops, and benefits from GPT-6's revised prompt-caching system that keeps cache hit rates high when reasoning effort is changed mid-conversation. Best-fit deployments include subagents, bulk classification, first-pass code changes, and any job run thousands of times where a failed attempt costs only cents.

DigitalOceanopenai-gpt-6-lunagpt-luna

Open provider page Provider docs

Quick Info

Provider: DigitalOcean
Model key: openai-gpt-6-luna
Release date: Sep 22, 2026
Last updated: Sep 22, 2026
Knowledge cutoff: 2026-05-18
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $0.10
Output token cost: $0.50

Limits

Input tokens: 922,000 tokens
Output tokens: 128,000 tokens
Context window: 1,050,000 tokens

Transparent token rates

Compare OpenAI GPT-6 Luna pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about OpenAI GPT-6 Luna

DigitalOcean

Coverage

OpenAI has launched two new AI models, GPT-6 Soul and GPT-6 Luna, expanding its GPT-6 family with a focus on reducing costs, improving accuracy, and lowering error rates compared with previous versions. The launch follows the September introduction of GPT-6 Astra, which OpenAI described as its most powerful model to date, designed for complex tasks including coding and computer use. Soul targets complex work like coding and advanced reasoning, while Luna is built for high-volume office tasks such as summarizing documents and answering quick questions. API pricing is $2 per million input tokens and $10 per million output tokens for Soul, and $0.10 and $0.50 for Luna, roughly a 50% cut from GPT-5.6. OpenAI reports Soul made about half as many errors as its predecessor, Luna beat version 5.6 by 5.4% on an automation test while costing 58% less per task, and the models arrive in ChatGPT Work, Codex, and the API, rolling out gradually to the main interface.

DigitalOcean

CoverageBenchmark

AI Catchup's technical analysis confirms GPT-6 Luna was released on September 22, 2026, positioned as the budget model under GPT-6 Astra and sibling to GPT-6 Sol. The model is available in the API as gpt-6-luna and is rolling out across ChatGPT Work, Codex, and GitHub Copilot. Luna features a 1.05M-token context window On pricing, Luna is $0.10 per million input tokens, $0.01 cached input, and $0.50 output, representing a 50% reduction from GPT-5.6 Luna's promotional pricing. For coding performance, AI Catchup reports Luna scores 66.6% on DeepSWE at max effort for approximately $0.22 per task, matching what Sol achieves at xhigh effo