OpenAI: GPT-6 Luna Pro

GPT-6 Luna Pro is a text model from OpenAI, released by the provider on 22 September 2026. It is the same underlying model as GPT-6 Luna, served with reasoning.mode set to pro, which is intended to produce higher-quality responses on complex tasks. If you already know the Luna family, the Pro variant is the configuration to reach for when answer quality matters more than the amount of reasoning work the model performs.

The model takes text, images and files as input, and returns text only. Its context window is 1,050,000 tokens, and the maximum output is 128,000 tokens. That combination of a very large input window and a large output ceiling means you can supply long documents, image inputs and file attachments in a single request and still leave room for a substantial generated response. Supported request parameters include max_tokens and max_completion_tokens for output length, response_format and structured_outputs for constraining the shape of the response, tools and tool_choice for function calling, seed for reproducibility, and verbosity for controlling how much the model says. Reasoning behaviour is exposed through the reasoning, reasoning_effort and include_reasoning parameters.

Practical use cases follow directly from those capabilities. Long-context work such as reviewing large codebases, contracts or research dossiers fits the 1,050,000 token window, and file and image inputs let you mix source documents with diagrams or screenshots in the same request. Structured outputs and response_format suit pipelines that need machine-readable results rather than prose, while tools and tool_choice support agents that call functions or external services. Because Pro mode is aimed at complex tasks, it is a reasonable default for multi-step analysis, careful drafting and problems where a weaker answer would cost more to fix than to generate.

You call the model through api14's OpenAI-compatible API using the model id openai/gpt-6-luna-pro. Point your existing OpenAI client at the api14 base URL and pass that id in the model field; the request and response shapes match what you already use. The parameters listed above are the ones the model supports, so you can set max_completion_tokens, response_format, tools, tool_choice, seed, verbosity and the reasoning-related fields in the same way you would with other OpenAI-compatible endpoints.

One thing to keep in mind is that pro mode spends far more reasoning effort than the standard Luna configuration, so it is not the cheapest or fastest choice for simple or high-volume work. For straightforward text generation, classification or short summaries, the non-pro GPT-6 Luna is likely the better fit. Reserve GPT-6 Luna Pro for requests where the extra reasoning translates into a materially better answer.