← All video models
Luma vs ByteDance
Ray3.2 vs Seedance 2.5
Two models line by line: prices from the official rates this site has confirmed, capabilities from the vendor's own page. Generation quality is not compared — we have no same-prompt samples we can publish, so anything said about it would be a guess.
The short answer
200 seconds of generation a month: 20 videos × 5 seconds each × 2 attempts on average.
- At this volume Ray3.2 costs less: $12 a month through Luma API, against ¥302 for Seedance 2.5 through Volcengine Ark.
- Channels this site can price: 20 for Ray3.2, 16 for Seedance 2.5.
Line by line
| Item | Ray3.2Luma | Seedance 2.5ByteDance |
|---|---|---|
| Basics | ||
| Developer | Luma | ByteDance |
| Status | Active | Active |
| Last confirmed | 2026-09-17 | 2026-09-17 |
| At your usage | ||
| Cheapest per month | approx.$12/mo Luma API · Pay as you go | approx.¥302/mo Volcengine Ark · Pay as you go |
| Pricing | ||
| Lowest subscription cost per second | $0.008/s Luma · Draft | $0.122/s Luma · 480p |
| Lowest API price per second | approx.$0.030/s Luma API · 540p · Not specified | approx.$0.100/s Volcengine Ark · 480p · Not specified |
| Official specs | ||
| Clip length | Not stated | 4–30 seconds |
| What it can do | Not stated | Text to video, Image to video, First/last frame, Reference to video, Edit video, Extend video |
| Highest resolution | Not stated | 1080p |
| Audio | Not stated | Generates its own audio |
| Specs (from the price lists) | ||
| Resolution | Draft · 540p · 720p · 1080p | 480p · 720p · 1080p |
| Audio | No audio | With audio |
| Billing | Per 5s clip, Per 10s clip | Per second, Per 5s clip |
| Where to buy | ||
| Subscription tools | Luma | Luma, Runway, InVideo AI, Pika, PixVerse |
| With an exact published rate | Luma | Luma, Pika |
| API | Luma API, Replicate | Runway API, fal, Volcengine Ark |
| Source | ||
| Official page | Official source page | Official source page |
The "at your usage" row uses this site's default volume; with your own numbers the answer can change. Compare at your own usage ↗