What Luma AI does
Luma AI is a private Redwood City, California–based AI company building multimodal “world model” capabilities for generating and operating in the physical world. Its consumer and developer surfaces are organized around a unified creative workflow: users create and revise image/video/audio through Luma’s apps and agents, while developers access the underlying generative models through Luma API.
More
Luma’s product strategy emphasizes (1) directable generation for production workflows (not just best-effort prompting), (2) continuity and iteration across creative steps using agentic tooling, and (3) multimodal reasoning integrated into model design. On the image side, Luma positions Uni-1 / Uni-1.1 as a unified reasoning-and-generation model; on the video side, Luma positions Ray (now Ray3.2 and related variants) as a reasoning video model with an API surface for frame-level creative control and production pipeline compatibility. Business model: Luma sells access through consumer subscription tiers (Plus/Pro/Ultra) and business/enterprise offerings, and it also provides usage-based developer access through its API (credits/tiers, plus a dedicated “Scale” tier described in its API documentation pages). Luma also highlights partnerships and enterprise go-to-market support (e.g., “Forward Deployed Creatives” for enterprise onboarding).