VIDEO
Luma Dream Machine on Elevence AI
Luma's Dream Machine — fast, cost-effective text-to-video with strong scene understanding.
Applies to web, iOS and Android
PROVIDER
Luma AI
MODALITY
video
PRICING
From $0.08 per second of generated video
Luma Dream Machine is one of the leading text-to-video and image-to-video models, with a sweet spot at cost-effectiveness and turnaround speed. It generates 5-second clips in under a minute, with quality that's competitive with Runway Gen-3 on many scene types and cheaper per second.
On Elevence AI you get Luma Dream Machine alongside Runway, Veo and Kling — switchable per scene. For high-volume preview work or storyboard pipelines, Luma is often the right default.
What Luma Dream Machine is great at
Best-in-class price per second
At ~$0.08 per second of generated video, Luma is roughly 2-3× cheaper than Runway Gen-3. For preview pipelines and high-volume work, the cost advantage compounds quickly.
Fast generation turnaround
Luma typically completes a 5-second clip in 30-90 seconds. Faster than Runway, much faster than Veo, which makes it well-suited to iterative storyboard refinement.
Strong scene understanding and consistency
Luma's Ray 2 model produces coherent scenes with consistent characters and lighting. Strong on physical-world scenarios — landscapes, urban environments, atmospheric outputs.
Best for
•
High-volume video preview and storyboard work
•
Cost-sensitive ad creative iteration
•
Atmospheric and landscape scenes
•
Quick turnaround for client review cycles
Limitations
Camera motion control less granular than Runway
Runway's motion-brush and explicit camera-direction prompts give finer control. Luma's outputs are excellent but the directorial precision is slightly behind Runway's hero-tier features.
Maximum clip duration
Like Runway, Luma outputs are typically 5 seconds per generation. For longer videos, use multi-scene sessions and stitch outputs together.
Code example
Drop-in OpenAI-compatible API. Swap the base URL and key, change the model field:
// Multi-scene Luma generation:
const session = await fetch('https://api.elevence.ai/api/v1/video/sessions', {
method: 'POST',
headers: { Authorization: `Bearer ${KEY}`, 'Content-Type': 'application/json' },
body: JSON.stringify({
model: 'luma-dream-machine-2',
scenes: [{ prompt: 'A coastal village at golden hour, slow drone pull-back' }],
}),
});
Frequently asked questions
Luma vs Runway — which should I pick?
For cost-sensitive high-volume work, default to Luma. For hero shots where directorial control and cinematic motion matter most, reserve Runway. Many teams use both — Luma for previews, Runway for finals.
How fast does Luma generate?
Typically 30-90 seconds per 5-second clip. Significantly faster than Veo (which can take 2-5 minutes) and comparable to or faster than Runway.
Does Luma support image-to-video?
Yes. Pass a starting image as the conditioning input and Luma animates from that frame. Useful for bringing brand visuals and existing assets to motion.
What's the maximum video length on Luma?
Single requests produce ~5-second clips. For longer videos, use Elevence AI's multi-scene workflow — generate multiple sequential clips with shared style and stitch them.
How does Luma pricing work on Elevence AI?
Pay per second of generated video, billed in credits. Live rate is shown before generation. No subscription or quota — the same credit pool covers chat, image, video and audio across 60+ models.
Related models
Try Luma Dream Machine on Elevence AI
Sign up free and access Luma Dream Machine alongside 60+ other models — one bill, pay-as-you-go.
Get started →