Luma UNI 1 Max: lock this still before you buy motion (1cr)

Luma UNI 1 Max is the stills row. Animate after approval. Do not generate the pack inside the video model.

Luma UNI 1 Max is Luma's Uni-1 (Max) image model: reasoning-first, plans intent before generating, tuned for art-directed cinematic stills, with optional reference images. Category is text-to-image. An image is not required. The catalog lists 1 credit. There is no duration, no video, no marked audio. Aspect ratios: 1:1, 16:9, 9:16, 4:3, 3:4. Features named: text-to-image, reasoning, cinematic, Luma.

One credit is how a team accidentally animates a draft. They skip this row because it looks like a thumbnail. Then they pay a video row to invent the pack.

Reasoning on a still is the cheap plan

"Plans intent before generating" is the catalog's claim. Use it. Write the art direction as if a DP had to light it: subject, lens, wardrobe, the one object that must be readable at phone size. UNI 1 Max is the 1-credit place to let that plan happen on a single frame. A video model will also "plan," and it will plan for seconds you have not approved, at a credit cost this row does not have.

Optional references are optional. If a face or a SKU must match, attach it. If you are exploring a look, do not attach seven cousins. One reference that is the look beats a pile that confuses the reasoner.

Pick the ratio on this row, not later. 9:16 stills become 9:16 motion. A 1:1 hero you later squeeze into a story is a crop, not a plan. The five listed ratios are the menu; there is no 21:9 and no 4K token on this slug. Do not brief 4K. Brief the frame.

The text-to-image tool is the launcher. Luma does not have a provider hub here. The stills map is best AI image generator. This page is UNI 1 Max at 1 credit.

Animate the approved frame, not a moodboard

A video generate that also "designs" the bottle will design a new bottle every take. That is the identity tax. Lock the bottle here. Print it. Get the nod. Then image-to-video. The motion model should inherit a contract, not a prompt.

1 credit is the listed cost. It is also the reason people skip approval: "it's only a still." The expensive behaviour is using a 1-credit model as a moodboard you never mark approved, then animating take four of an unchosen look. Pick a still. Name it locked. Then buy motion.

There is no audio to write. There is no 4s / 6s / 8s. If you are arguing about soundtrack or length, you have already left this row. Come back only if the picture in the clip is the problem — and then you are editing or replacing the still, not "fixing UNI in the timeline."

A plate, not a clip

Not a video model. Not a lipsync plate-maker that moves. It can be the plate for those jobs. That is a two-step workflow, and the first step is this 1-credit still.

Not an editor. If the still is almost right, an edit-image row points at the object. UNI 1 Max is a generate. Regenerating to move a logo a centimetre is how reasoning-first becomes reasoning-again.

Not a duration bargain. People compare 1 credit to video rows and conclude they should "just do video." That comparison is backwards. The still is the cheap insurance. The video row is the motion you buy after the insurance pays out.

Would you send the PNG?

If you would not send this still to a client without a clip, you are not allowed to animate it.

If yes, you have a locked still. Take it to image-to-video. Do not re-prompt the pack as text in the video box "to save a step." You would be spending the 1-credit work twice, once at video prices.

If you only have a sentence and you need motion today, you can still run UNI 1 Max first. One credit. One frame. Then the clip. That order is the whole article.

FAQ

Is Luma UNI 1 Max a video model?

No. Text-to-image, 1 credit, cinematic stills, optional references, ratios 1:1 / 16:9 / 9:16 / 4:3 / 3:4. Motion is a later row. Lock the still here.

Do I have to upload a reference?

No. requires_image is false. References are optional and useful when identity must match. Exploring a look can be prompt-only. Shipping a pack should not be.

Why not design the product in the video generate?

Because every video take re-rolls identity. UNI 1 Max is the 1-credit reasoner for a single approved frame. Animate the approved frame. Do not ask a motion model to invent the SKU.

Can I get 4K or audio from this row?

The catalog does not list a max resolution or audio on UNI 1 Max. It is a still. If you need 4K motion with a soundtrack, that is a video row after this still exists — and only after you would ship the still on its own.