Max Resolution
720p · 1080p native workflows · external upscaler supports 4K delivery
AI / Video Generation & Editing / Luma AI
Luma Dream Machine
Luma AI’s production-grade Ray3.2 video system for cinematic generation, multi-keyframe direction, performance-preserving editing, reframing and HDR/EXR workflows.
Text-to-VideoImage-to-VideoMulti-KeyframeVideo EditingMotion TransferHDR / EXRReframeDeveloper API
Max Duration
Up to 20 seconds at 1080p for Ray3.2 production workflows
Generation Modes
Text-to-Video · Image-to-Video · Multi-Keyframe · Extend · Modify Video V2 · Reframe
Native Audio
Ray3.2 is visual-first; Luma App integrates separate ElevenLabs SFX/Music/v3 audio models
About Luma Dream Machine
Ray3.2 is Luma AI’s current video model, released in June 2026. It targets professional production with text-to-video, image-to-video, multi-keyframe direction, Modify Video V2, Reframe, motion transfer, HDR and 16-bit EXR workflows, with clips up to 20 seconds at 1080p.
Specifications
Developer & Integration
1 specifications +
| API Access | Ray3.14 is currently web-only; not available on Luma API |
Overview
4 specifications +
| Main Specialization | Production-grade generative video, video editing and reframing with frame-level control |
| Target Audience | Filmmakers, VFX/post teams, agencies, enterprise studios, creators and developers |
| Best For | Cinematic T2V/I2V, multi-keyframe direction, performance-preserving edits, motion transfer, HDR/EXR and campaign reframing |
| Current Model | Ray3.2 — released June 9, 2026 |
Pricing
2 specifications +
| Free Plan | Free/limited Dream Machine access varies by current plan and model availability |
| Pricing model | Subscription + generation credits |
Generation
8 specifications +
| Text-to-Video | Ray3.2 text-to-video through Luma App and Agents API |
| Image-to-Video | Start/end-frame I2V plus multi-keyframe generation |
