Luma AI unveils Ray3, a generative video model with HDR support

Luma AI has introduced Ray3, a video generation model that it claims is the first capable of producing studio-quality HDR (High Dynamic Range) videos.

Ray3 supports 10-, 12-, and even 16-bit color depth and can export results as EXR files, making it easier to integrate with professional editing and color-grading workflows. The model can also convert standard SDR videos into HDR.

"Reasoning" capability and visual control

One of Ray3’s most notable features is what Luma AI calls its "reasoning" ability - the capacity to think in both images and language, follow complex instructions, and assess its own outputs. According to the company, Ray3 can critique early drafts and iterate until it meets a given quality bar.

Video: Luma AI

This reasoning ability also enables visual control: users can guide movement, objects, or camera framing by sketching directly on images instead of typing text prompts.

Video: Luma AI

Ray3 introduces a Draft Mode that generates rough video previews up to five times faster and at lower cost. These drafts can then be upgraded to full 4K HDR "Hi-Fi" quality in a second pass.

The company highlights Ray3’s higher fidelity and realism, including improvements in rendering crowd scenes, light interactions and reflections, motion blur, and character consistency. The model is already available for free through Luma AI’s Dream Machine app.

Luma AI frames Ray3 as a creative tool "built to tell stories" and has released several short films made with the technology. Unlike Google’s Veo 3, however, it cannot generate audio natively.