Ray 3.2
by Luma AI model
Luma's current video model, which the company describes as putting "every frame under your control", holding "continuity across every cut" and finishing in cinematic quality — continuity across cuts being the hard part and the thing worth testing it on. It sits inside Luma Agents, which spans video, image, audio and text, alongside Uni-1 for holding a brand's style consistent across generations and Luma Skills for workflow automation, added in June 2026.
Visit siteGood at
- Text to Video 4 / 5