GPT-6 Luna API

Technical specifications of GPT-6 Luna

These specifications are from OpenAI's current GPT-6 Luna model documentation.

What is GPT-6 Luna?

GPT-6 Luna is designed for applications where a large number of model calls need to be handled efficiently. OpenAI describes it as its most efficient model for focused, high-volume tasks, while its broader GPT-6 guidance recommends Luna for efficient, repeatable work at scale.

Main features of GPT-6 Luna

• Very large context: The 1.05M-token context window supports long documents, large code contexts, and multi-stage application workflows.
• Configurable reasoning: Six reasoning-effort settings let developers trade reasoning depth against latency and token usage. medium is the documented default.
• Multimodal input: Luna accepts text and images, while its model output is text. Audio and video are not supported as model modalities.
• Production tool use: The Responses API exposes web search, file search, code interpreter, hosted shell, computer use, MCP, tool search, and other tools for Luna.
• Structured application integration: Function calling and structured outputs are supported, making Luna suitable for classification, extraction, routing, automation, and other machine-readable workflows.
• Efficient economics: OpenAI lists standard pricing of $0.10 per 1M input tokens and $0.50 per 1M output tokens, with separate rates for cached input, long-context requests, Batch/Flex, and Fast mode.

Benchmark performance of GPT-6 Luna

OpenAI's current public model page provides the model specifications and capabilities but does not present a conventional benchmark table for Luna on that page. Independent Artificial Analysis reports an Intelligence Index of 37 for its highest-reasoning GPT-6 Luna configuration, with lower scores for lower reasoning settings; this is an independent evaluation rather than an OpenAI-published benchmark.

Because benchmark methodology and reasoning settings materially affect results, these figures should not be treated as a universal ranking.

GPT-6 Luna vs GPT-6 Sol vs GPT-6 Astra

The key distinction is workload positioning, not context size: all three currently have a 1.05M-token context window and 128K maximum output. OpenAI describes Astra as its most capable model, Sol as the model for demanding reasoning, and Luna as the efficiency-oriented option for high-volume work.

Representative use cases

1. High-volume classification and extraction — structured outputs and low per-token pricing fit repetitive workloads.
2. Customer-support and chat systems — Luna's positioning emphasizes focused, repeatable calls at scale.
3. Long-document processing — the 1.05M-token context can accommodate unusually large source collections.
4. Multimodal document workflows — image input enables applications that combine text with screenshots, diagrams, or scanned material.
5. Agentic automation — Responses API tools such as file search, code interpreter, hosted shell, computer use, MCP, and tool search support multi-step workflows.
6. Cost-sensitive coding assistance — Luna supports reasoning, function calling, structured outputs, and coding-related tools, while Sol is positioned for more demanding coding and agentic work.

Limitations

GPT-6 Luna does not support fine-tuning, and its native model modalities exclude audio and video. OpenAI also documents an important endpoint distinction: in Chat Completions, function calling for Luna requires reasoning_effort to be set to none; OpenAI recommends the Responses API for built-in tools and function calling.

How to Access GPT-6 Luna API with CometAPI

GPT-6 Luna is designed for efficient, high-volume AI workloads. If you want to access GPT-6 Luna through CometAPI, the basic integration pattern is straightforward: create a CometAPI API key, configure an OpenAI-compatible client to use CometAPI's API base URL, and select the gpt-6-luna model ID when the model is enabled for your account.

CometAPI provides an OpenAI-compatible API gateway, allowing developers to use familiar OpenAI SDKs while accessing models from multiple providers through one API.

Step 1: Create a CometAPI Account

First, sign up for or log in to your CometAPI account.

After logging in, open the API Keys section in the CometAPI dashboard and create a new API key. CometAPI states that newly created keys use an sk- prefix.

Keep the API key on the server side and store it in an environment variable rather than hard-coding it into application source code.

Step 2: Configure the CometAPI Base URL

CometAPI's current OpenAI-compatible base URL is:

https://api.cometapi.com/v1

This means applications using the OpenAI Python or Node.js SDK can generally switch to CometAPI by changing the API key and base_url.

Step 3: Select GPT-6 Luna

The CometAPI model catalog identifies GPT-6 Luna with the model ID:

gpt-6-luna

Before deploying it in production, verify that gpt-6-luna is currently enabled and that the live catalog provides an endpoint for your account. CometAPI's current GPT-6 Luna comparison page lists the model ID but does not currently populate its endpoint and pricing fields.

Why Use CometAPI for GPT-6 Luna?

The main advantage is a unified OpenAI-compatible interface. Instead of implementing a separate SDK and authentication system for every supported model provider, developers can use the CometAPI base URL and switch the model value when moving between supported models. CometAPI currently advertises access to more than 500 models through this approach.

This can be useful when an application needs to experiment with different models, maintain a multi-model architecture, or migrate between providers without rewriting its entire API integration.