Production teams wanting tight control over AI video get 16 keyframes and EXR export, but there is no free tier and 1080p eats credits.
Kenning's take
What it does
Ray3.2 is an update to Luma's Ray3 video models, built with studios, agencies and game makers. It lets users place up to 16 keyframes in one clip, carry an actor's body and facial performance into new footage, reframe shots and export 16-bit HDR and EXR files for colour grading. It is available in the Luma app on paid plans and, for the first time with its full controls, through Luma's API.
- Up to 16 keyframes in a single clip
- Modify Video V2 for clips up to 20 seconds at 1080p
- Performance and facial expression transfer
- Reframe for any aspect ratio
- Native 16-bit HDR with EXR export
Who it suits. Film, advertising and game production teams who need precise control over AI video.
Key facts
- Category
- AI models
- Part of
- Luma Dream Machine
- Made by
- Luma AI
- Based in
- United States
- Released
- 9 June 2026
- How to get it
- in the Luma app on paid plans, and through the Luma API
- Works on
- Web, iOS, API
- Free plan
- none listed on the pricing page
Pricing
| Plan | Price | Billed |
|---|---|---|
| Plus | $30 | month |
| Pro | $90 | month |
| Ultra | $300 | month |
| Team | $contact sales | custom |
Reviews and write-ups
FAQ
When was Ray3.2 released?
Luma announced Ray3.2 on 9 June 2026.
Is Ray3.2 available through an API?
Yes. Luma says the full set of Ray controls is available through its API for the first time.
How many keyframes can Ray3.2 use?
Up to 16 keyframes inside one clip.
Facts from the maker's own pages and the sources listed. Checked 2026-10-11.
Similar tools
Ray3.14
Luma video model with native 1080p output, faster and cheaper than the original Ray3
Uni-1
Luma's image model that reasons about a request in language before generating or editing a picture
ACE-Step 1.5
Open-source music generation model that writes full songs with lyrics and runs on ordinary consumer GPUs.
AlphaGenome Atlas
Google DeepMind database of AI-predicted effects for every possible single-letter change in human DNA.