OpenAI GPT

OpenAI's language models, connected to Nova and her knowledge base: a more accurate brief, a storyboard from a precise mock-up, a method that remains yours.

OpenAI GPT refers to OpenAI's language models, the GPT-6 family, available through the API. At Axiris, they do not replace Nova: Nova can rely on a local model or on an OpenAI model to speak, but the knowledge, the method and the memory of your project remain those of Axiris. What a GPT brings on top: a fine reading of an image or a mock-up, and a quality storyboard from a precise composition. This link is in preparation; a first bridge exists in internal testing.

OpenAI's language models

OpenAI documents a family of language models, the GPT-6 generation, which reads text and images and answers in text. Its documentation breaks it down into several models, from the most capable for demanding work to the most efficient for high-volume tasks. This page does not list them by version: they change often, and OpenAI's documentation remains the reference.

Nova, and the model that makes her speak

Nova is the Axiris assistant for image and video. What makes her who she is is not the model that makes her speak: it is her knowledge base, specialised in the image professions, the studio's method, and the memory of your project. Nova

Nova can therefore rely on a local model or on an OpenAI model, depending on what you ask. In both cases, the knowledge remains that of Axiris. You do not choose the technical route; you say what you want to produce, and the confidentiality level you set for the project decides what may leave. Your data does not leave, unless you decide it does

What a GPT brings on top

A storyboard from a precise mock-up. A GPT writes good storyboards. It writes better ones when given an image: a mock-up built in the studio, even in low definition, is enough for it to imagine the scenes, the framings, the sequence of shots. With Nova having prepared the brief, the storyboard that comes out is ready to go to video, with a cloud model such as Seedance or in the studio's video engine. This is the scenario we are preparing. Box Director

A fine reading of the image. Describing a reference, naming what is wrong in a render, rewording a client brief into production instructions: a GPT does this well, and Nova knows when to call on it.

The image, alongside. OpenAI's image generation has its own page: GPT Image 2.5.

What you need to know

An OpenAI model is a remote model: what you send it goes to OpenAI, under its terms. Local AI model or API?

For a business that wants nothing to leave, Nova works with a local model, and her knowledge base never leaves Axiris. Nova for business

Where the link stands

The link between Nova and OpenAI's models is in preparation. A first conversation bridge exists in the studio, in internal testing; it is not open to accounts. The opening will be announced in the news.

Official source

OpenAI, model documentation

Contact request: a first conversation, free and with no commitment. It is by listening to you that we propose what suits you.