Robots Atlas>ROBOTS ATLAS
GPT-6 Luna
AI Models›GPT

GPT-6 Luna

GPT-6 Luna (gpt-6-luna) · Family: GPT
The budget model of OpenAI's GPT-6 family, released on 22 September 2026: $0.10 per million input tokens and $0.50 per million output — half the price of GPT-5.6 Luna.
✓ Active✓ Public accessLLMMultimodalTool-using model📁 GPT
Release date
22 September 2026
🏢OpenAIProducer
Access:APIHostedDeployment:☁ Cloud

Overview

GPT-6 Luna is the cheapest model in OpenAI's GPT-6 family, released on 22 September 2026 together with the mid-tier GPT-6 Sol — roughly three weeks after the flagship GPT-6 Astra. It targets high-volume use cases where the unit cost of a request matters more than the model's peak capability.

The model runs under the API identifier gpt-6-luna and costs $0.10 per million input tokens and $0.50 per million output tokens. Against its predecessor GPT-5.6 Luna, priced at $0.20 and $1.20 per million, that is roughly a halving of price — and alongside a stated capability upgrade rather than at the expense of quality.

Luna closes the three-tier GPT-6 line-up from below: above it sits GPT-6 Sol at $2 and $10 per million tokens, and at the top GPT-6 Astra at $10 and $50. The price gap between Luna and Astra is therefore a hundredfold on input tokens. OpenAI did not disclose Luna's context window size or detailed benchmark results in the launch material.

Classification
LLMMultimodalTool-using model
Family: GPT
Access & deployment
APIHosted
Cloud
Weights: Closed
Key parameters
📥 Input: text, image, documents

Technical specification

Modalities
⬇ Input
textimagedocuments
⬆ Output
textcodestructured_data

Capabilities and applications

Native model capabilities
Reasoning
The model's ability to reason logically and solve complex problems.
Category: reasoning
Coding
Generating, analysing and modifying code in many programming languages. Covers writing functions, debugging, refactoring, code review, and creating tests. Measured by benchmarks such as HumanEval and SWE-bench.
Category: coding
Tool use
The model's ability to call external functions, APIs and tools during a conversation: calculator, search engine, code editor, database. The model decides when and how to use a tool and interprets its result.
Category: planning
Function Calling
Category: planning
Structured output
Producing data in structured formats such as JSON.
Category: structured_generation
Image understanding
Analysing and interpreting the content of images.
Category: vision
Multilingual
Competence in many natural languages (from a few to over a hundred): understanding, generation, translation, and code-switching within a single conversation. Frontier models support a wide range of languages with comparable quality.
Category: language
Classification
Assigning an observation to one of predefined classes (binary or multi-class). Output: class label and optionally probabilities.
Category: other
Instruction following
Precisely following instructions contained in the prompt: response format, length, style, constraints (e.g. 'reply in six words'). GPT-5.1 significantly improved this capability compared to GPT-5.
Category: language

Pricing

Sources and related pages