Luma Uni-1.1 Guide: Features, Price and Prompts

Luma Uni-1.1 is Luma's reasoning image model, built for realistic people and characters that keep their look across reference images: in Flick it makes an image from a prompt and up to 9 references for 10 credits at Standard.

An ink drawing of a hand holding up a lit oil lantern beside a small framed portrait of a woman

Luma describes Uni-1.1 as a model where reasoning and drawing happen in one network rather than two systems stitched together. Search results for it show API docs and price lists, but almost no pictures of what it makes. For this Luma Uni-1.1 guide we made nine images in Flick, from a noodle-shop owner carried into two new scenes to a Japanese manga page, an Arabic film poster, a storyboard and the same premiere shot at Standard and Max. Below are its features, prices, the exact prompts and what came back.

Describe your first Luma Uni-1.1 shot

Write the character, the place and the light in the box, and it opens a new Flick project where you can pick Luma Uni-1.1 and make the image.

How do you use Luma Uni-1.1 in Flick?

Add an image node, pick Luma Uni-1.1, connect any character or style references and @-mention them in the prompt, choose Standard or Max, and run it.
  1. 1Pick Luma Uni-1.1

    Add an image node on a Flick canvas, choose Luma Uni-1.1 in the model list, and set one of its five frame shapes: 16:9, 9:16, 1:1, 3:2 or 2:3.

  2. 2Connect references

    Connect up to 9 images of a character, a place or a style and @-mention each one in the prompt, so the model knows which picture is which.

  3. 3Choose the quality

    Leave Quality on Standard for drafts, and switch to Max for a final frame, which doubles the price and, in our test, drew sharper detail and more of the prompt.

What is Luma Uni-1.1?

Luma Uni-1.1 is an image generation and editing model from Luma AI that reasons about a prompt before it draws, and reads up to nine reference images to keep a person, place or style.
Luma Uni-1.1 in Flick, from the app's model config on November 13, 2026, Luma's announcement and our own test files
Luma Uni-1.1 in Flick
MakerLuma AI
Ways to use itImage node, Text to Image and Edit Image
Price at Standard10 credits
Price at Max20 credits
Each reference image+1 credit
Reference imagesUp to 9
Frame shapes16:9, 9:16, 1:1, 3:2 and 2:3
Default 16:9 image, measured2672 × 1504
Picker estimateabout 30 seconds
PlanFree
Where these figures come from

Prices, settings, reference limits, the frame shapes and the plan are read from Flick's model config; the image size is measured on our test files, and the time is the model picker's estimate, not a measurement.

  • Luma says Uni-1.1 renders text natively in non-Latin scripts, including Chinese, Japanese and Arabic, and holds continuity across several reference images, in its Uni-1.1 API announcement.
  • Luma calls the first version, Uni-1, "a multimodal reasoning model that can generate pixels", on its Uni-1 page.
  • Luma reports its own rankings on human preference tests; we have not checked them and do not repeat them here.

How much does Luma Uni-1.1 cost in Flick?

Luma Uni-1.1 costs 10 credits an image at Standard and 20 at Max, plus 1 credit for each reference image you connect.
Luma Uni-1.1 prices in Flick, in credits, from the app's model config on November 13, 2026
QualityNo referenceOne referenceEach extra reference
Standard (default)1011+1
Max2021+1
  • Free to try: Luma Uni-1.1 is on Flick's Free plan, and a new account gets 100 credits.
  • References stay cheap: with all 9 references connected at Standard, an image is 19 credits.
  • The node shows the total: the price on the run button updates as you change quality or connect references.

What did Luma Uni-1.1 make in our tests?

In our tests Luma Uni-1.1 drew convincing film lighting and wrote English, Japanese and Arabic correctly, but it missed some layout instructions, such as the reading order of a manga page.

Each card shows the prompt exactly as it was run, at Standard quality, with no reference image or style.

A 1940s woman in a dark office turning in shock toward a framed photo of a man on her desk, blinds casting shadows

A reveal told without words

Asked for the moment a detective realises who the killer is, it put a framed photo of a man on her own desk in the foreground and turned her face to it, though the frame cuts his face in half.

A film still from the moment just after the last line of this scene: "The detective finally understands who the killer is, and it is the one person she trusted." Show her face and what she is looking at, so a viewer gets it without words. Noir lighting, a 1940s Los Angeles office at night, cinematic.
Flick team · Luma Uni-1.1 tests
A road movie poster with Arabic and English titles over two women in a dusty red car on an empty desert highway at sunset

An Arabic film poster

It wrote "الطريق الطويل", "THE LONG ROAD" and the tagline exactly, with nothing else on the poster, but drew a car seen head on rather than a pickup truck.

A movie poster for a desert road film. Title in large Arabic calligraphy: "الطريق الطويل" with "THE LONG ROAD" in English underneath. Two sisters in a dusty pickup truck on an empty highway at sunset, long shadows, heat haze. Tagline at the bottom: "Home is 900 kilometres away." Nothing else written.
Flick team · Luma Uni-1.1 tests
A black and white manga page of a girl finding an old film camera in an attic and raising it to take a photo

A manga page in Japanese

All three Japanese lines came back letter for letter, but the second landed one panel early, it added a fourth balloon, and the sound effect lost its first character.

A black and white manga page with five panels, screentone shading, read right to left. A teenage girl finds an old film camera in her grandfather's attic. Speech balloons in Japanese: panel 1 "おじいちゃんの箱だ…", panel 3 "これ、まだ動くの?", panel 5 "撮ってみよう!". A sound effect "カシャッ" in the last panel. Dynamic angles, expressive faces.
Flick team · Luma Uni-1.1 tests
A four-panel ink and grey wash storyboard of a girl rowing to a whale stranded on a sandbar beside a lighthouse

A four-panel storyboard

It typed all four labels in the right panels but swapped each hyphen for a long dash, and drew the whale as a fish in the aerial panel.

A film storyboard sheet with four panels in a 2 by 2 grid, black ink and grey wash on white paper. Story: a lighthouse keeper's daughter rows out to a stranded whale at dawn. Under each panel a typed label: "1. WIDE - whale on the sandbar", "2. MEDIUM - she pushes off in a rowboat", "3. CLOSE - her hand on the whale's eye", "4. AERIAL - the tide comes in". The same girl and boat in every panel.
Flick team · Luma Uni-1.1 tests

Does Luma Uni-1.1 keep a character the same across references?

Partly: with the portrait connected as a reference, two new scenes kept her face shape, grey-streaked hair and taped hands, but dropped the eyebrow scar and did not make her younger when asked.

We made a portrait first, then connected it to two new prompts, each for 11 credits with one reference.

A Korean woman in her forties with grey-streaked hair and taped hands leaning on a noodle shop counter under a bare bulb
The reference: a retired boxer at her noodle shop, with a cut above her left eyebrowFlick team · Luma Uni-1.1 tests · Luma Uni-1.1
Photoreal cinematic portrait of a 45-year-old Korean woman, a retired boxer who now runs a noodle shop: short grey-streaked hair, a faded scar through her left eyebrow, strong hands wrapped in old athletic tape, a white T-shirt under a navy apron. She leans on the counter of her small shop at closing time, a single bare bulb above her, steam rising from a pot behind her, 50mm lens, shallow depth of field, Kodak Portra colours, no text.
The same woman in a boxing ring with red gloves raised, sweat on her face, a crowd blurred behind the ropes

Reference 1: back in the ring

Her face and grey streak carried over and the gloves, ring and crowd matched the prompt, but she looks no younger than the reference, and the scar is gone.

The woman in the reference image, same face, scar and grey-streaked hair, seventeen years earlier at age 28: in a boxing ring under hard top light, sweat on her face, red gloves raised, a crowd blurred behind the ropes. Cinematic, 35mm film grain.
Flick team · Luma Uni-1.1 tests
The same woman in profile under a clear umbrella on a rainy Seoul overpass at night, neon reflected on the road

Reference 2: home in the rain

It kept the taped hands from the portrait without being asked and set her under a clear umbrella in neon rain, though she stands still rather than walking.

The woman in the reference image, same face, scar and hair, walking home alone along a Seoul overpass in the rain at night, holding a clear umbrella, neon signs reflected on the wet road, wide shot from the side, her face in profile. Cinematic, anamorphic.
Flick team · Luma Uni-1.1 tests

Is Luma Uni-1.1 Max worth double the price?

In our one test, yes: Max drew the Art Deco front, velvet ropes, falling rain and a crisp marquee that Standard missed or blurred, for 20 credits against 10.
Cinematic wide shot of a film premiere outside an Art Deco cinema in 1920s Paris at night: a crowd in evening dress behind velvet ropes, press photographers with flash powder, a marquee reading "LE DERNIER PHARE", rain on the cobblestones reflecting the lights, a vintage car pulling up at the kerb. 35mm film, warm tungsten and cold blue.

Standard

Standard: a stone cinema front at an angle with LE DERNIER PHARE in lit letters, a crowd behind rope and an open vintage car
Flick team · Luma Uni-1.1 tests · Luma Uni-1.1

Max

Max: a symmetrical Art Deco cinema with a bulb-lit LE DERNIER PHARE marquee, red velvet ropes and photographers in the rain
Flick team · Luma Uni-1.1 tests · Luma Uni-1.1
  • Standard spelled "LE DERNIER PHARE" correctly on an angled sign, but the building reads as a plain stone front, the ropes as thin cord, and the figures on the left are smeared.
  • Max framed the cinema head on with a sunburst Art Deco window, a bulb-lit marquee, red velvet ropes, top hats and rain streaks in the light.
  • Both put a flash of powder on the left and a vintage car at the kerb, and neither added text beyond the marquee.

How do you write a good Luma Uni-1.1 prompt?

Describe the moment and the light as a cinematographer would, quote every word to appear, and name the details a reference must carry, since Uni-1.1 kept what it was told and guessed the rest.
  1. 1. Describe the moment

    Write what has just happened, not only what is in frame, because our detective prompt gave no props and still got a framed photo placed to tell the twist.

  2. 2. Quote every word

    Put titles and dialogue in quotes and say "nothing else written", which kept the Arabic poster to its three lines with no invented text.

  3. 3. Write dashes as words

    Avoid a hyphen inside storyboard labels, or accept a long dash, since Uni-1.1 changed all four hyphens in our storyboard.

  4. 4. Name what must carry over

    List the face, hair, scars and costume a reference must keep, and check small marks by eye, because our boxer kept her hair and hands but lost the scar.

  5. 5. State panel order plainly

    Number panels by position ("top right", "bottom left") rather than by reading direction, because the manga page put a line one panel early.

  6. 6. Save Max for finals

    Draft at Standard and rerun the keeper at Max, which sharpened the marquee and followed the Art Deco and rain details Standard missed.

How does Luma Uni-1.1 compare with GPT Image 2.6?

At default settings in Flick, GPT Image 2.6 costs 5 credits against 10 for Luma Uni-1.1, while Uni-1.1 returns a larger image and has a quicker picker estimate.
Luma Uni-1.1 and GPT Image 2.6 in Flick, from the app's model config on November 13, 2026; image sizes measured on our test files
Luma Uni-1.1GPT Image 2.6
Credits at default105
Default 16:9 image, measured2672 × 15041536 × 864
Reference imagesUp to 9Up to 10
Picker estimateabout 30 secondsabout 2 minutes
PlanFreeFree

For both models run on the same prompts, see Luma Uni-1.1 vs GPT Image 2.6.

Luma Uni-1.1 FAQ

What is Luma Uni-1.1?

Luma Uni-1.1 is Luma AI's image generation and editing model. Luma describes it as a reasoning model that thinks through a prompt and draws in one network, and it reads up to nine reference images.

How much does Luma Uni-1.1 cost?

In Flick, Luma Uni-1.1 costs 10 credits an image at Standard quality and 20 at Max, plus 1 credit for each reference image connected.

Is Luma Uni-1.1 free?

Luma Uni-1.1 runs on Flick's Free plan, and each image uses credits. A new Flick account gets 100 credits, enough for several Luma Uni-1.1 images at Standard.

What is the difference between Uni-1.1 and Uni-1.1 Max?

Max is Uni-1.1's higher quality setting, at double the price. In our premiere test, Max drew sharper sign lettering and followed more of the prompt's details, such as the Art Deco front and the rain.

How many reference images does Luma Uni-1.1 take?

Up to 9. In Flick you connect them to the image node and @-mention each one in the prompt, and each adds 1 credit to the price.

Can Luma Uni-1.1 write Japanese or Arabic text?

Yes, in our tests. It wrote three Japanese speech balloons and an Arabic film title, "الطريق الطويل", correctly, and Luma says it renders Chinese, Japanese and Arabic natively.

What aspect ratios does Luma Uni-1.1 support in Flick?

Five: 16:9, 9:16, 1:1, 3:2 and 2:3. A default 16:9 image measured 2672 × 1504 in our tests, and a 2:3 poster 1632 × 2464.

Sources

More on AI images in Flick

Make your first Luma Uni-1.1 image in Flick

Pick Luma Uni-1.1 on an image node, paste a prompt from this guide, and keep the result beside the rest of your film.

Make a Luma Uni-1.1 image in Flick

Try for free, no credit card required.

Ask AI about this page:ChatGPTPerplexityClaude