Uni-1.1 - Luma AI's New Generation Image Generation Model

What is Uni-1.1

Uni-1.1 is a new generationimage generationmodel and API service launched by Luma AI, adopting a decoder-only autoregressive Transformer architecture that integrates text reasoning and pixel generation into a unified process. The model supports joint input of up to 9 reference images, sentence-level image editing, complex layout generation, and multilingual text rendering, ranking third globally on the Arena.ai blind test leaderboard. The API offers two modes: pay-as-you-go billing and reserved throughput, with a minimum cost of approximately $0.04 per image, targeting enterprise-level scenarios such as advertising, e-commerce, and content creation.

Main Features of Uni-1.1

• Text-to-Image: Generates high-quality images based on text prompts, capable of outputting complex layouts containing a dozen or so types of layout elements such as headers, navigation, advertisements, and body text in a single pass.
• Image Editing: Performs multi-round editing based on sentence-level instructions, preserving unmentioned elements by default, enabling iterative visuals much like editing a document.
• Multi-Reference Image Fusion: A single call supports joint input of up to 9 reference images, performing semantic-level fusion of brand logos, products, real people, characters, etc. as model-level hard constraints.
• Spatial and Pose Control: Supports precise control such as rotation, viewpoint switching, and spatial relationship adjustment, maintaining subject identity and texture without loss.
• Multilingual Rendering: Supports high-quality text generation for non-Latin characters such as Chinese and Arabic, meeting global content needs.

Technical Principles of Uni-1.1

• Unified Autoregressive Architecture: Adopts a decoder-only autoregressive Transformer, where text tokens and image tokens share the same sequence, enabling cross-modal joint reasoning.
• Integrated Reasoning and Generation: The model performs cross-modal reasoning before generating pixels; constraints such as composition, space, and brand consistency are solved at the structural level, rather than translating first and then drawing.
• Dual-Endpoint API Design: Provides a Reasoning endpoint (deconstructs instructions, plans composition, locks brand/character/product constraints) and a Generation endpoint (completes pixel rendering based on reasoning results).
• Reference Image Hard Constraint Mechanism: Passes multiple reference images as model-level hard constraints to ensure visual identity remains consistent across all channels and versions.

How to Use Uni-1.1

• Register an Account: Visit the Luma AI developer platform official website (https://platform.lumalabs.ai）注册并登录账号。
• Obtain a Key: Create a project in the developer console and obtain an API Key.
• Choose a Billing Mode: Choose the Build plan (pay-as-you-go, suitable for flexible calls) or the Scale plan (reserved throughput, minimum order of 8 units, suitable for large-scale production) based on usage.
• Call the Reasoning Endpoint: Send text instructions and reference images, letting the model deconstruct requirements, plan composition, and lock brand/character constraints.
• Call the Generation Endpoint: Complete pixel rendering based on reasoning results and obtain the final generated image.
• Integrate the SDK: Integrate the API into existing workflows through the officially provided Python, JavaScript, TypeScript, Go, or CLI SDK.
• Upload Reference Images: Pass up to 9 reference images as hard constraints in the request to ensure the output is consistent with the brand's visual identity.
• Iterative Editing: Use sentence-level editing instructions to make multiple rounds of adjustments to generated results, gradually optimizing until satisfactory.

Key Information and Usage Requirements of Uni-1.1

• Product Name：Luma Uni-1.1 / Uni-1.1-Max
• Publisher: Luma AI (core research team of fewer than 15 people)
• Release Date: May 6, 2026
• Product Positioning: Enterprise-grade AI image generation model and API service
• Technical Architecture: decoder-only autoregressive Transformer (integrated reasoning and generation)
• Leaderboard Ranking: No. 3 globally on Arena.ai (behind OpenAI gpt-image-2 and Google nano-banana-2)
• Price Range: Build plan text-to-image $0.0404–$0.1000 (2048px); Scale plan monthly fee $2,100–$3,800/unit
• Enterprise Customers: Adidas, Mazda, Publicis Groupe, Serviceplan, Envato, Comfy, Krea, etc.
• SDK Support：Python、JavaScript、TypeScript、Go、CLI
• Core Team: Song Jiaming (Jiaming Song, author of DDIM), Shen Bokui (William Shen, CVPR Best Paper)

Core Advantages of Uni-1.1

• World's No. 3 Generation Quality: Ranked No. 3 globally in Arena.ai user blind-test ELO ratings, behind only OpenAI gpt-image-2 and Google nano-banana-2.
• Extreme Cost-Effectiveness: A single 2K-resolution image costs as little as $0.0404, with both price and latency less than half of comparable top-tier models.
• Enterprise-Grade Consistency: Through hard constraints from reference images and sentence-level editing, it solves the pain points of traditional models such as character deformation, brand color drift, and inconsistent styles across markets.
• Complex Tasks Completed in a Single Pass: It can generate a complete, readable news website page or a full set of advertising campaign assets in one go, without the need for multi-module stitching.

Comparison of Uni-1.1 with Similar Competitors

Uni-1.1 application scenarios

• Advertising localization: Rapidly expand key visuals into multilingual and multi-regional versions, lock in brand elements through reference images, and greatly shorten production cycles.
• E-commerce product visualization: Generate consistent product images in real time based on product photos, fabric samples, and scene references, replacing traditional shooting and template-based workflows.
• Character and IP consistency: Provide character consistency guarantees across scenes, poses, and lighting for game promotional art, comics, and film/TV pre-production.
• Brand content pipeline: Integrate into enterprise content production systems to enable batch generation and style unification of visual assets across markets.
• Creative prototype design: Combine hand-drawn sketches with material references to quickly generate realistic product concept images and 3D clothing renders.