What is a unified model?

A model that handles text-to-image generation and editing of an existing image through the same engine, instead of two modes that behave differently. The benefit is continuity of look between the generation and the retouch.

What is the difference between Uni-1 and Uni-1 Max?

Max is the top tier of the same family. Both cover the same five formats and the same modes; you iterate on Uni-1 and switch to Max for the final version.

Does Luma do video on Beemm Vision?

Yes, with Luma Ray 3.2: a text becomes a 5 or 10 second shot, a starting image a 5 second shot, at 540p, 720p or 1080p. Ray 3.2 does not generate sound.

Do I need a Luma account?

No. Beemm Vision calls the provider for you and bills in credits on a single balance shared across the whole catalogue.

Model · Luma

Luma Uni-1 for images, Ray 3.2 for video

Luma Uni-1 is a unified model: generating from text and reworking an existing image go through the same engine, not two tools that behave differently. In practice an edit looks like the original generation rather than a graft. Ray 3.2, Luma’s video family, completes the catalogue: a text or a starting image becomes a moving shot.

What Luma brings

One engine for both moves

Generating and editing are the same operation for this model. The practical consequence: the render does not change character when you edit — which is exactly the problem with chains where the retouch comes from a different model than the generation.

Two tiers, one simple choice

Uni-1 to iterate, Uni-1 Max when the render goes to production. Switching happens in the picker, with no change to the prompt or the settings.

The five formats that matter

Square, 16:9, 9:16, 4:3 and 3:4 for Uni-1 — the real coverage of social display and layout. The format is chosen before generating, not by cropping after.

Video with Ray 3.2

Ray 3.2 turns a text into a 5 or 10 second shot, or a starting image — with an optional end image — into a 5 second shot. Output at 540p, 720p or 1080p, six formats from 21:9 to 9:16. The model does not generate sound.

At home in a chain

On a canvas, Luma is one brick among others: it can take an image from another model, rework it, then hand off to a crop or animation node.

The Luma engines available here

Uni-1 and Uni-1 Max exist in both text-to-image and image-to-image form; Ray 3.2 handles video.

Luma Uni-1 MaxThe top tier of the family, png or jpeg output, five ratios.
Luma Uni-1The standard release, same format coverage.
Luma Ray 3.2Text to video (5 or 10 s) or image to video (5 s, optional end image), 540p to 1080p, six formats, no audio.

What Beemm Vision adds

A unified model is compared like any other: on your subject, not on a leaderboard.

Run one brief through Luma, Seedream and FLUX and keep the one that serves your image.
Chain into an edit, or animate the Uni-1 image with Ray 3.2, in the same graph.
Replay a workflow over a list to produce a consistent series.
Ship the graph as an app for the team.
One credit balance, and no Luma API key to manage.

Frequently asked questions

What is a unified model?: A model that handles text-to-image generation and editing of an existing image through the same engine, instead of two modes that behave differently. The benefit is continuity of look between the generation and the retouch.
What is the difference between Uni-1 and Uni-1 Max?: Max is the top tier of the same family. Both cover the same five formats and the same modes; you iterate on Uni-1 and switch to Max for the final version.
Does Luma do video on Beemm Vision?: Yes, with Luma Ray 3.2: a text becomes a 5 or 10 second shot, a starting image a 5 second shot, at 540p, 720p or 1080p. Ray 3.2 does not generate sound.
Do I need a Luma account?: No. Beemm Vision calls the provider for you and bills in credits on a single balance shared across the whole catalogue.

Try Luma Uni-1 and Ray 3.2

The button opens Generate with Luma already selected. The full catalogue is available from the free plan.

Open the studio See the other models How credits work