Ray3 is an intelligent video model designed to tell stories.
Ray3 is capable of thinking and reasoning in visuals and offers state-of-the-art physics and consistency. In a world’s first, Ray3 generates videos in 16-bit High Dynamic Range color, bringing generative video to pro studio pipelines. The all-new Draft Mode enables you to explore many more ideas, much faster, and tell better stories than ever before.
Modify Video
Reimagine any video. Shoot it in post with director-grade control over style, character, and setting. Restyle expressive performances, swap entire worlds, or redesign the frame to your vision. Shoot once. Shape infinitely.
_
.
.
.
.
.
.
.
.
.
.
Luma's mission is to build multimodal general intelligence that can generate, understand, and operate in the
physical world
AI for the next billion people will not look like LLMs. Computers won’t be point and click. Tomorrow’s systems will be creative, immersive, and multimodally interactive.
Text alone is not enough to get there. The current industry approach of scaling LLMs on limited text data has hit a wall. At Luma we are doing foundational research and building systems that can train efficiently on rich multimodal data. This is the path to useful general intelligence.
We are training world models that jointly learn from video, audio, and language — similar to how the human brain learns. This will enable models to see, hear and reason about the world so they can effectively collaborate with us, help us communicate better, entertain us, and one day, operate alongside us in the real world.
We are a small stellar team of researchers, engineers, and product thinkers dedicated to this singular mission. If that's you, we offer an opportunity to do your life's work and solve one of the most important problems of our time in a collegiate and collaborative environment. Join us.
Platform
Foundations for a new era of creativity and human expression
Worldbuilding is the foundation of storytelling. AI that is able to understand context, intent, and imagination, and gives you the power to manipulate anything is the key to unlocking this new era of creative expression.
We are building powerful and collaborative products for creatives working on video across media, entertainment, marketing, and advertising industries. These products are built in concert with Luma's intelligent multimodal video, audio and reasoning models and make rich and controllable worldbuilding accessible.
Ideate, visualize, create videos, and share your dreams with the world, using our most powerful image and video AI models. Available now on iOS and the Web. Learn more.
Research
We are focused on foundational research and systems engineering to build multimodal general intelligence
We are a team of some of the world's best researchers and engineers from leading universities and AI labs with a track record of inventing foundational technologies like DDIM, the VAE Bottleneck, joint video-image-audio-text generation, Neural Radiance Fields, advanced neural compression, and realtime generation.
We are building the mathematical foundations and systems necessary for training and scaling multimodal world models with deep understanding of physics, causality, creative thinking, and reasoning abilities.
Team & Community
Everything we can imagine should be real
We're bringing together pioneering researchers, product thinkers, cracked engineers, and a community of daring creatives to shape the future of world building and multimodal intelligence.
Join us in shaping this revolution!
We're a small, high-achieving team building multimodal general intelligence If you share our vision of AI that transcends current boundaries to enable extraordinary capabilities and human-AI collaboration, we'd love to talk.
Explore the possibilities unlocked through world building. Find tutorials, best practices, and inspiring examples from our community of millions of creatives. Learn how others are using Dream Machine, and the Luma API to transform their creative process across design, film, education, and beyond.