Ray 3.2

Overview

Ray 3.2 is Luma's text-to-video model for cinematic short clips. With just a text prompt, it generates 5-second or 10-second clips at 540p, 720p, or 1080p, and supports six aspect ratios: 16:9, 9:16, 1:1, 4:3, 3:4, and 21:9. It responds reliably to structured, specific shot and subject descriptions, and performs especially well on atmospheric, environment-focused shots. It suits developers and creative teams that need to quickly produce concept films, storyboard previews, and social short videos, and can be integrated into generation workflows by calling it through the unified HiAPI.ai API.

Key Highlights

• Shot-level description control — It follows structured, specific shot blocking and subject descriptions closely, while results degrade under vague, open-ended instructions.
• Cinematic visual style tuning — Its visual style and physical realism lean cinematic, making it well suited to atmospheric empty shots and environment footage.
• Multiple aspect ratios and resolutions — Output supports six ratios spanning landscape, portrait, square, and widescreen, at resolutions up to 1080p and durations of 5 or 10 seconds.

Capabilities

Direct text-to-video generation: Enter a prompt of up to 6,000 characters to produce a short clip with no source footage required, ideal for quickly validating shot ideas from scratch.

Selectable duration and resolution: Clip duration can be 5 or 10 seconds, and resolution covers 540p, 720p, and 1080p, so you can choose specs separately for draft and final-output stages.

Six aspect ratios in one model: It supports 16:9, 9:16, 1:1, 4:3, 3:4, and 21:9, so the same prompt can directly produce landscape, portrait, or widescreen versions to fit different placements.

Cinematic and atmospheric shots: Its visual style and physical realism lean cinematic, and it is relatively stable on environment shots such as atmospheric empty shots and scene backgrounds.

Use Cases

• Film or advertising teams creating concept films and storyboard previews — Describe shot blocking and scene atmosphere in text to produce 5- to 10-second clips as early-stage visual references for creative and client discussions.
• Content teams generating atmospheric empty shots and background assets — Describe lighting, environment, and mood to produce environment shot assets for transitions, background screens, and atmosphere building in video editing.
• Marketers quickly producing social short videos — Choose portrait or landscape aspect ratios by placement and directly generate 5- to 10-second clips suited to each platform for social media and ad placements.
• Developers generating shot assets in batches — Submit generation tasks in batches by prompt through the unified HiAPI.ai API and feed shot assets into your own content production or asset library workflows.

Limitations and Selection Advice

Clips are capped at 10 seconds and 1080p, so it is not suited to long takes or direct high-resolution master output. Third-party evaluations show that temporal consistency and motion quality are not its strongest areas, and shots with continuous motion or multiple interacting subjects require multiple generation rounds and filtering. Results depend on structured, specific prompts, and performance degrades under vague or open-ended instructions. If your needs involve long-form narrative, fine-grained motion control, or image-driven generation, evaluate other models or combine this with an image-to-video pipeline.