Luma Ray3.2 🧠 Model Paid
Luma's HDR video model with up to 16 keyframes in a 20-second 1080p clip
- Maker
- Luma AI
- Latest release
- 2026-06
- Weights
- Closed
- Licence
- –
- Apps offering it
- 10
Where you can use Luma Ray3.2
Apps in this directory that let you generate with Luma Ray3.2, from their own documentation.
Good for
About Luma Ray3.2
What it is
Ray is Luma AI's video model line. Ray3 introduced reasoning-driven generation and native HDR. Ray3.14 (January 2026) made it faster, and Ray3.2 (June 2026) adds frame-level direction. Ray runs inside Luma's own app, which is now organised around 'Luma Agents', and through Luma's API.
Key features
- Clips up to 20 seconds at 1080p
- Up to 16 keyframes inside one clip for precise direction
- Native HDR generation for grading-friendly output
- Strong video-to-video: character reference, modify and restyle
- Launch materials do not mention native audio. Luma's app offers ElevenLabs audio models separately
Where you can use it
Luma's app (Dream Machine / Luma Agents) and the Luma API. Adobe Firefly Boards lists Ray2, Ray3 and Ray3.14 as partner models. ComfyUI has Luma text- and image-to-video partner nodes, and Krea lists Ray 2 as a legacy model.
Pricing and rights
Luma's pricing page lists no free tier. Plus costs $30 a month ($25 billed yearly) for 10,000 credits, Pro $90 and Ultra $300. All three include commercial use and access to Luma's own and third-party models (Veo, Kling, Seedance, MiniMax). We found no published API rate for Ray3.2.
Who it is for
Filmmakers and VFX artists who need exact timing and HDR plates. Creators who direct by keyframes rather than by prompt alone.
Verdict
Ray3.2's keyframe control and HDR output are distinctive and useful for professional work. There is no native audio, no free tier, and a higher entry price than most rivals.
Pros
- Up to 16 keyframes per clip
- Native HDR output
- 20-second 1080p clips
- Commercial use on all paid plans
Cons
- No native audio mentioned
- No free tier; $30/month entry
- No public API rate found
Similar AI models
All video models →Veo 3.1 🧠 ModelPaid
Google DeepMind's text- and image-to-video model with native sound, up to 4K
Kling 3.0 🧠 ModelFreemium
Kuaishou's video model with native multilingual audio, lip sync and 15-second clips
Seedance 2.5 🧠 ModelPaid
ByteDance's video model: 30-second clips with audio, many references and local edits
MiniMax H3 (Hailuo 3.0) 🧠 ModelFreemium
MiniMax's open-weight video model: native 2K, stereo sound, 15-second clips
Runway Gen-4.5 🧠 ModelFreemium
Runway's own video model, now with native audio and one-minute multi-shot scenes
Wan 3.0 🧠 ModelOpen source
Alibaba's video family: open-weight Wan 2.x and the 30-second Wan 3.0 API
More from Luma AI
Luma (Dream Machine) Paid
Luma's Ray3 video studio, now also running Veo, Kling, Seedance and GPT Image
🔥 Save up to 20% with yearly billing (Plus $25/mo)