Luma UNI 1 Max is a stills row
UNI 1 Max is Luma’s 1-credit image SKU: reasoning-first cinematic stills. It is not a clip, a duration, or a talking generate.
Luma UNI 1 Max is an image SKU. Category is text-to-image. One credit. No duration, no marked audio, no video out. The catalog description is a reasoning-first stills model that plans intent before generating, tuned for art-directed cinematic results, with optional reference images. People still send it a 6-second brief because the brand name Luma is associated with motion. The catalog does not care about the brand association. This row draws a frame.
That is a routing rule, not an approval-then-animate sermon. If you needed motion, you will pick a video row after you admit this card cannot emit one. If you needed lettering, you will pick a typesetting stills row instead, because cinematic reasoning is not a UI engine.
Luma in this catalog is an image SKU
UNI 1 Max’s listed features are text-to-image, reasoning, cinematic, Luma. Aspects: 1:1, 16:9, 9:16, 4:3, 3:4. An image is not required. References are optional. There is no 21:9. There is no 4K token on this slug. Do not brief 4K. Brief the frame.
The 1-credit sticker is how a team treats it as a thumbnail they can ignore. Ignore it and you will spend a video rate inventing the look this row would have planned for a credit. That is not an argument for animating every UNI still. It is an argument for not skipping the stills row because you “came to Luma for video.”
On Versely, click this slug when the deliverable is a picture: a key art frame, a mood board hero you will actually pick, a cinematic still that has to hold a lens and a wardrobe. Click a video slug when the deliverable is seconds. Do not infer a clip from the vendor.
Text to image is the launcher. The stills ranking is best AI image generator. This page is only UNI 1 Max as a catalog type.
What you can attach: optional references for a face or an object that must recur. What you cannot attach: a driving clip, a wav, a duration chip. Those inputs belong on motion-control, lipsync, or text-to-video rows. Pasting a video URL into a stills prompt does nothing useful.
Reasoning-first is a stills brief
“Plans intent before generating” is the catalog’s claim. Write a brief a DP could light: subject, lens, wardrobe, time of day, the one object that must read at phone size, palette. UNI 1 Max will spend the pass on whether that brief coheres, then put pixels down. A vibe word is not a brief. “Cinematic” is already in the feature list; repeating it does not add a camera.
Useful prompts name:
- Who or what is in frame, at what scale.
- The light (window, tungsten, overcast), not “moody.”
- The lens feel (35mm close, 85mm portrait), not “bokeh bokeh.”
- What must not appear (extra hands, extra logos, extra skyline).
Useless prompts name camera moves. There is no timeline. “Slow dolly in” on a stills row is a wasted clause. If the still is later animated, describe the camera on the video row.
Optional references personalize. One reference that is the look beats a pile of cousins. Seven near-duplicate bottles confuse a reasoner. requires_image is false, so exploring a look can be prompt-only. Shipping a pack should not be: attach the SKU you already locked on a lettering or pack-shot row.
Pick the ratio here. 9:16 stills become 9:16 motion if you later animate. A 1:1 hero squeezed into a story is a crop of a reasoned composition. The five listed ratios are the menu.
Optional references, not a driving clip
A reference image on UNI 1 Max is “look like this person or this object.” It is not motion control. It will not copy a dance. It will not viseme a line. Teams that used Luma tools for motion still reach for UNI 1 Max with a performance clip attached. That clip is the wrong noun.
If the next job is motion:
- Camera or product turn from this still: image-to-video on a video row, start frame = this file.
- Face performance from a driving take: a motion-control SKU, not UNI 1 Max.
- Talking: a lipsync SKU after a plate and a track exist.
UNI 1 Max can be that plate. That is two steps. The first step is this 1-credit still. The second step is a different slug. Collapsing them because “Luma” feels like one product is how you skip the cheap plan.
Not an editor. If the still is almost right, HiDream O1 Image Edit or another edit-image row points at the object. Regenerating UNI 1 Max to move a glass a centimetre is reasoning-again, a new drawing, even at 1 credit.
Where UNI 1 Max loses
Lettering. Posters, pack type, UI chrome, podcast titles. Ideogram, Seedream, Recraft, GPT Image 2 — those stills rows letter. UNI 1 Max is cinematic reasoning. A pretty frame with cousin spelling is a miss for a cover.
Dense boards. Infographics, comparison tables, six-panel carousels. Seedream 5.0 Pro is the dense-layout row. UNI 1 Max will art-direct a single hero, not a spreadsheet of claims.
Strict prompt obedience on a catalog object. GPT Image 2 lists strong prompt adherence and 4K. UNI 1 Max lists cinematic bent and no 4K token. If the brief is “this exact SKU, this exact composition, 4K,” you may be on the wrong stills row.
Anything with seconds. No durations. No native mix. No talking flag. If the brief names 6 seconds in the first sentence, leave.
Print at sizes this record does not promise. There is no listed max resolution. Do not write “4K key art” against UNI 1 Max. Generate the look, then decide whether another stills row or an upscaler owns delivery pixels.
The lock-then-animate checklist for this SKU already exists elsewhere on the site. This page is the type error: UNI 1 Max is a stills row in a vendor family people remember as video. Send it a stills brief. Spend the credit on a frame you would actually pick.
When the brief is art-directed cinematic stills at 1 credit, sign up and run UNI 1 Max as image, not as a clip you hope it will become.
FAQ
Can I set 6 seconds on Luma UNI 1 Max?
No. The catalog lists no durations. Six seconds is a video row. UNI 1 Max draws one frame at 1 credit.
Do I have to upload a reference to use it?
No. References are optional. Prompt-only is valid for look exploration. Attach a reference when a face or SKU must match. Do not attach a driving video and expect motion transfer.
Why would I use this instead of a 2-credit photoreal stills row?
You are paying for reasoning-first, cinematic stills with those five aspects, not for a generic sample. If you need lettering, dense layout, or a listed 4K ceiling, pick the row that publishes that job. If you need a planned cinematic frame, this is the 1-credit SKU.