Video models
From MiniMax H3's consistent characters to Seedance's big motion and Veo's native sound. Pay per second, switch any time.
MiniMax
Cinematic video with consistent characters and real voices
from $0.007/s
ByteDance
ByteDance's newest video model, multi-reference and long takes
from $0.116/s
ByteDance
Proven ByteDance video model at 720p
from $0.161/s
ByteDance
The budget Seedance
from $0.089/s
Google's video model with native sound
from $0.10 / video
| Model | Maker | Specs | From |
|---|---|---|---|
| MiniMax H3 | MiniMax | 480p · 768p · 1080p · 1–15 s · Up to 9 reference images · Up to 3 voice references · Lip sync · 2 aspect ratios | $0.007/s |
| Seedance 2.5 | ByteDance | 720p · 480p · 4–30 s · Up to 30 reference images · 3 aspect ratios | $0.116/s |
| Seedance 2.0 | ByteDance | 720p · 4–15 s · Up to 9 reference images · 3 aspect ratios | $0.161/s |
| Seedance 2.0 Mini | ByteDance | 720p · 4–15 s · Up to 9 reference images · 3 aspect ratios | $0.089/s |
| Veo 3.1 | Fast · Components · Pro · Up to 3 reference images · 2 aspect ratios | $0.10 / video |
One balance for all of them. Failed generations are refunded automatically.