400 credits for $4.99 USD, 1,200 credits for $14.99 USD, 3,000 credits for $29.99 USD, 7,200 credits for $59.99 USD. A 5-second 480P video costs about $1.25. Credits are one-time purchases and never expire.
Key features
Pyramid-JiT (P-JiT) architecture with 32×32 token reduction
4× more aggressive token reduction than Linum v2
Converges in 11× fewer samples than Linum v2
2B-parameter model with frozen Qwen text encoder
Pixel-space diffusion with multi-resolution readouts
PixelREPA alignment with DINOv3-L
Curriculum learning over noise schedule
Motion Transfer mode that rebuilds the entire video around original motion
Object Swap mode to change selected targets only
Replace multiple people with up to 10 reference photos per video
Swap characters, outfits, products, or objects
Optional prompt for specific instructions
Output quality options from 480P to 1080P
Cost preview before video generation
Support for videos 4 to 30 seconds long up to 100 MB
Preserves original motion, timing and camera movement
What makes it different
Low-cost training: full 2B model trained for just $21K in compute
Controls baked into model training rather than post-hoc fine-tuning
Pixel-space compression without separate VAE bottleneck
Apache 2.0 open release with full model weights and code
No subscription required - pay per video with one-time credit packs
Credits never expire
Transparent pricing shown before generation starts
Linum
Genjutsu