Luma is an AI video generation model by Luma AI, whose Dream Machine platform runs the Ray3 family — the current Ray3.2 (June 2026) is a reasoning-driven, director-grade model with frame-level control, native 16-bit HDR, and best-in-class camera and 3D-spatial motion.
Built for filmmakers and creative teams, it turns briefs into cinematic clips — with precise camera direction, 3D spatial consistency, and a native HDR pipeline that drops straight into professional post-production.
Cinematic clips, ads, animation, camera-driven and 3D-consistent shots, previs
Film and marketing teams that need precise camera direction and production-grade control
The camera-first video model — 3D-consistent motion, reasoning, and a native HDR pipeline.
Rooted in Luma's NeRF origins, Dream Machine's biggest differentiator is spatial consistency — the scene holds together in 3D as the camera moves. Dolly, pan, tilt, follow, orbit, and zoom are all directable with precision, so complex camera moves land instead of warping the scene.
Ray3 was the first reasoning video model: it evaluates and refines the shot internally before delivering it, which improves physical accuracy and prompt adherence. The model interprets a shot brief rather than approximating it, so structured direction resolves as intended.
Ray3.2 introduces frame-level control, giving production teams precise command over how action moves and changes from start to finish. With this update, creators aren't just prompting — they're directing, frame by frame, which is what makes output usable as a real production asset.
Native 16-bit HDR tone-mapping with EXR export means output that drops into professional color and post-production workflows, not just a flat social clip. Paired with EXR, it's ready for broadcast and film-grade delivery.
The Ray3.14 update delivers native 1080p — no upscaling — at 4× faster generation and 3× lower cost than base Ray3, typically completing in under two minutes. Fast iteration makes wide creative exploration practical instead of a waiting game.
Luma AI fits wherever teams need camera-driven, production-grade video.
Add a node, direct the camera, and generate cinematic clips with production-grade control.
Add a Luma node to the canvas and describe your idea in a text prompt — or upload a starting frame to animate, or existing footage to modify. Set keyframes (start and end frames) if you want to control how the shot begins and ends.
Direct the camera explicitly — dolly, pan, orbit — and set resolution and duration. Ray3.14 renders native 1080p in under two minutes; Ray3.2 adds frame-level control for precise action. Choose HDR/EXR output when you need a production-grade pipeline.
Use Modify Video to restyle or adjust, chain the node with an image model for the first frame or an upscaler for 4K delivery, then export with commercial rights on your plan and move to the next shot.
Real generated videos with the exact prompts — copy any prompt and run it yourself.
One subscription across 30+ AI models — no per-tool credit balances or separate signups. Credit cost per generation is shown live in the node before you run it.
500 weekly credits to test the model and see the output quality. No credit card required.
45,000 monthly credits with cheaper per-credit pricing, commercial rights, and video download.
90,000 monthly credits, up to 10 seats, shared workspace, centralized billing, and priority support.
210,000 monthly credits, API and integration support, unlimited seats, and dedicated management.
Full specs for the Luma model family.
How Luma compares with other AI video models in the Phygital+ catalog.
Chain image, video, upscaling, and audio into one repeatable workflow.
Design the first frame with an image model, then animate it with Luma in one workflow.
Generate with Luma, then upscale to 4K delivery resolution for film, ads, or broadcast.
Add narration, sound effects, or a music track with a voice or audio node to finish the clip.
Write the script and shot list with a text model, then feed it straight into video generation.
Everything you need to know about Luma in Phygital+.
Luma is Luma AI's AI video generation platform, Dream Machine, powered by the Ray3 model family. The current Ray3.2 (June 2026) is a reasoning-driven, director-grade model that turns text, images, or existing video into cinematic clips — with native 1080p, native 16-bit HDR, frame-level control, and the camera direction and 3D spatial realism Luma is known for.
Usage is credit-based and included in every Phygital+ plan — from the free plan up to Enterprise. The exact credit cost per generation is shown live in the node before you run it and varies with duration and resolution. This is separate from Luma's own Dream Machine pricing, which offers a free draft tier and paid plans from around $30/month.
Outputs created on any paid plan (Starter and up) come with commercial-use rights. Note that on Luma directly, commercial use is gated to the Standard/paid tiers — the free tier is watermarked with no commercial license. Luma's terms also prohibit generating deepfakes of identifiable real people without consent. Check the model card for any provider-specific limits.
Luma is made by Luma AI, a San Francisco lab founded in 2021, originally on NeRF (Neural Radiance Fields) technology for 3D scene reconstruction. It launched Dream Machine in June 2024 and has iterated through the Ray series to the current Ray3.2.
Ray3.2, announced June 9, 2026, is the newest — built for frame-by-frame direction with continued HDR/EXR strength. The Ray3.14 update (January 26, 2026) added native 1080p, 4× faster generation, and 3× lower cost. Both build on Ray3, the first reasoning video model.
Ray3.14 and later generate native 1080p without upscaling, with a 4K upscale available for delivery. On Luma directly, 1080p output is gated to paid plans.
Sound effects can be enabled as a premium add-on, but Luma does not have full native audio generation the way some rivals do. For dialogue and rich soundtracks, a dedicated voice or audio node — which you can chain in Phygital+ — is usually the better fit.
Modify Video is Luma's video-to-video feature: transform existing footage up to 18 seconds — restyle it, swap subjects or objects, or apply a look — while holding motion consistency across frames. It turns Dream Machine into an AI-assisted post-production tool.
Ray3.14 (January 2026) focuses on native 1080p, 4× speed, and 3× lower cost, but doesn't support Character Reference or HDR/EXR — you fall back to base Ray3 for those. Ray3.2 (June 2026) adds frame-level control for precise direction and keeps HDR/EXR as a strength. Pick Ray3.14 for fast, cost-efficient 1080p; Ray3.2 for director-grade control and HDR pipelines.
Yes. Luma offers a Dream Machine API with per-clip pricing. In Phygital+, you can run Luma as a node on any plan, and API and integration support is available on the Enterprise plan.
Browse the full catalog of 30+ AI models available in Phygital+.
Join 100+ teams using Phygital+ – every model in one workspace
Try Phygital+ free
