Interpret intent before pixels
Luma positions Uni-1.1 as a reasoning model that resolves creative intent before generation. Write prompts with scene logic, subject relationships, lighting, camera language, and the details that must stay consistent.
Luma uni-1.1 reasoning visual generation and editing
Luma describes Uni-1.1 as a production-grade reasoning model for the Luma API: it interprets creative intent before generating, then supports both new visual creation and natural-language modification. Use this workspace to write structured prompts, guide results with references, choose the framing, preview outputs, and download final visuals.
uni-1.1 creation workspace
Use uni-1.1 for reasoning-guided prompts, reference editing, previews, and downloads.
Source reference
Describe your idea
0/5000Model
Aspect Ratio
Output Format
Recent Jobs
Visual examples
A compact gallery of Luma uni-1.1 output directions. Use the images to choose a visual style before generating with Luma uni-1.1.
Features
Luma positions Uni-1.1 as a reasoning model that resolves creative intent before generation. Write prompts with scene logic, subject relationships, lighting, camera language, and the details that must stay consistent.
Use the create workflow for a new composition, or upload a source reference and describe the edit in plain language. This maps to Luma guidance around Create and Modify use cases.
Luma guidance recommends telling the model what each reference is for: style, character, composition, color palette, lighting, texture, or mood. Clear reference roles reduce guessing.
Control the generation mode, prompt, aspect ratio, output format, and repeatable iterations without exposing unnecessary model plumbing. The page keeps attention on choices that affect the final result.
Inspect the generated result, load it back as the next edit source when you need another pass, or download the final asset from the result panel.
Model strengths
Luma describes Uni-1.1 as the production-grade reasoning model behind the Luma API. The important SEO point is not just “AI generator”; it is a model that interprets intent before rendering, then follows multi-constraint creative direction.
The official API framing separates two jobs: Generate for text prompts and reference-guided creation, and Modify for natural-language edits such as changing lighting, swapping backgrounds, applying a reference aesthetic, or making localized changes.
Use Luma uni-1.1 when a prompt needs reference grounding, composition control, product or character continuity, clear edit instructions, seed-based iteration, and multiple aspect ratios rather than a generic one-shot prompt.
How to use it
Start with the core Luma uni-1.1 decision: are you creating a new result, or modifying an existing source?
01
Use text prompts for a new composition. Use source editing when the output should remain a version of the uploaded reference.
02
Describe the subject, layout, lighting, texture, camera angle, background, and any reference role such as style, character, composition, lighting, or mood.
03
Submit the job, wait for the result, preview the output, then use the download action on the completed result.
Use cases
Luma uni-1.1 fits work where the prompt is a creative brief: a subject, references, constraints, what changes, and what must stay recognizable.
Compare
Use this as a practical selection guide, not a lab benchmark. Luma emphasizes Uni-1.1 for reasoning-driven generation, reference grounding, and plain-language edits, so it is strongest when the task behaves like a creative brief rather than a short decorative prompt.
Luma describes Uni-1.1 as a reasoning model that interprets intent before generating. That makes it better suited to detailed creative briefs, multi-reference direction, scene reasoning, and natural-language edits than a short style-tag prompt.
Yes. Use the create workflow for a new composition. Describe the subject, scene, style, lighting, framing, reference roles, and output intent, then generate and preview the result.
Yes. Use the modify workflow when the output should remain a version of your source. Write what should change, what should stay untouched, and what style, setting, light, or background you want.
Write prompts like a compact creative brief: subject, composition, lighting, environment, material details, camera language, constraints, and reference roles. For edits, explicitly say which parts must remain unchanged.
It can help when references are labeled clearly, such as character, product, composition, color palette, or lighting. Upload the source, state the role, and describe the identity, shape, layout, or brand details that should remain recognizable.
Check brand rules, licensing needs, recognizable people or trademarks, text accuracy, product shape, and any platform policy that applies to your campaign.
Open the Luma uni-1.1 workspace, keep uni-1.1 selected, generate one result, and download it when the job finishes.