--- title: "UNI-1 | Less Artificial. More Intelligent." description: "Uni-1 is a multimodal reasoning model that can generate pixels. Built on Unified Intelligence, Uni-1 understands intention, responds to direction, and thinks with you. Read technical report to learn more." canonical: "https://lumalabs.ai/uni-1" source: "https://lumalabs.ai/uni-1.md" --- # UNI-1 | Less Artificial. More Intelligent. Uni-1 is a multimodal reasoning model that can generate pixels. Built on Unified Intelligence, Uni-1 understands intention, responds to direction, and thinks with you. Read technical report to [learn more](https://lumalabs.ai/uni-1/tech-specs). [Try Uni-1 Now](https://app.lumalabs.ai/) · [Build with Uni-1.1 API](https://lumalabs.ai/api) ## What Uni-1 can create [Explore more](https://lumalabs.ai/uni-1/showcase) ## Intelligent Common-sense scene completion, spatial reasoning, and plausibility-driven transformation. ### More examples ## Directable Reference-guided generation with source-grounded controls. ### More examples ## Cultured Culture-aware visual generation across aesthetics, memes, and manga. ### More examples ## Evaluations Uni-1 ranks first in human preference Elo for Overall, Style & Editing, and Reference-Based Generation, and second in Text-to-Image. ## Pricing ### Image Generation Pricing | | Per 1M tokens in USD | | --- | --- | | Input price (text) | $0.50 | | Input price (images) | $1.20 | | Output price (text and thinking) | $3.00 | | Output price (images) | $45.45 | ### Equivalent per-image price* | | Equivalent per-image price* | | --- | --- | | Text to Image (2048px) | $0.0909 | | Image edit / i2i (2048px) | $0.0933 | | Multi-ref, 1 img (2048px) | $0.0933 | | Multi-ref, 2 imgs (2048px) | $0.0957 | | Multi-ref, 8 imgs (2048px) | $0.1101 | *Per-image prices based on billing token counts. Each image (input or output) = 2,000 billing tokens at current settings. All prices in USD. [Build with Uni-1.1 API](https://platform.lumalabs.ai/) ## Get API Access Join the waitlist to get early API access to Uni-1. We'll notify you when API is available. [Build with Uni-1.1 API](https://platform.lumalabs.ai/) ## UNI-1 FAQs ### General ### What is Uni-1? Uni-1 is Luma’s first unified understanding and generation model. Capable of any ask involving images and text, it’s a powerful tool for editing and generation. Having been trained on the full gamut of creative tasks, Uni-1 helps you turn ideas into finished work. ### What can I use Uni-1 for? - **Image generation** - Starting from a text prompt, Uni-1 can create an endless range of image genres and styles. - **Image editing **- Given an input image and modification instructions, Uni-1 can make highly directable edits. - **Reference **- Uni-1 can generate an output image using up to 9 images as reference. ### How does Uni-1 stand out among other image models? Uni-1 stands out through its intelligence, directability, and cultural awareness. It reasons through your prompts rather than simply matching keywords, gives you precise control over camera angles, lighting, and composition, and has deep knowledge in a variety of creative tasks. Compared to other image models, we consider Uni-1 best-in-class at several capabilities, such as: - **Cinematic image generation** - **Manga and webtoon generation** - **Style transfer** - **Style reference** - **Novel view synthesis** Scroll up to see demos of each of these use cases. ### What is unified intelligence? Unified models internally represent different modalities in one interleaved sequence, allowing them to understand and generate multiple forms of information such as text, code, images, video, speech, and audio. This differs from specialized models that are built to handle a single input and output format, such as traditional text-to-image generators. ### What makes Uni-1 "intelligent"? Uni-1 is Luma’s first model built in this unified line of research. Its formulation allows it to use understanding signals to improve generation, and generation training to strengthen fine-grained visual understanding. To learn more about Uni-1 under the hood, visit our model report. As a creative tool, Uni-1 understands your intent by combining its wide range of world knowledge with a careful analysis of your input, and generates images highly informed by this reasoning. The unified architecture enables flexible prompting and extremely diverse outputs with less of an “AI look”. ### Where can I find Uni-1 in the product? Uni-1 is available in the Luma AI app under the image generation section. Simply select Uni-1 as your model when creating a new image generation project. ### What’s coming next? We’ll continue upgrading Uni-1 by accepting more reference images, adding multi-output and multi-turn capabilities, and improving identity preservation, visual quality, and instruction-following. In parallel, we plan on growing our unified model suite to modalities beyond image and text. ### Usage ### How do I use Uni-1? To use Uni-1, sign in to your Luma account, navigate to the image generation section, select Uni-1 as your model, and enter your prompt. You can provide detailed direction including camera angles, lighting, style references, and more for precise control. ### How do I prompt Uni-1? Uni-1 is a highly directable model that’s trained to take any sort of instruction. You can give it short prompts, long prompts, structured JSONs or word salad. When words aren’t enough, you can doodle on an input image to direct your edit, or make a collage to direct exactly where each element should go. Try bringing in new images and referencing their composition, mood, or art style - the directions are unlimited. ### Can I control camera and lighting in my generations? Yes! Uni-1 offers precise directability. You can specify camera angles, heights, and distances, control light intensity, angle, and shadow strength, adjust color temperature, saturation, and contrast, and define motion and style preferences. ### Does Uni-1 support temporal sequences? Yes, Uni-1 can maintain consistency across time while evolving scenes through coherent motion and event progression. This makes it ideal for creating keyframe sequences and storyboards. ### Can I use reference images? Yes, you can use up to 9 reference images to guide your shot. ### What aspect ratios are available? Available aspect ratios are 3:1, 2:1, 16:9, 3:2, 1:1, 2:3, 9:16, 1:2, and 1:3. ### Where does Uni-1 fit in my workflow? Uni-1 can support multiple stages of the creative process, from early ideas to final outputs, and works alongside your existing tools. As we further develop our unified model suite, we hope to minimize the need to switch between tools by bringing more of the creative process into one place. ## Available only from Luma [Try Uni-1 for free](https://app.lumalabs.ai/)