Open-source generative AI models for image, video, audio and 3D creation
Generative video models for animation
Stage
Not given
Not given
Founded
2022
Not given
Based in
GB
US
Pricing model
Freemium
Not given
Pricing
Free for non-commercial use and for commercial use up to $1M annual revenue under Stability AI Community License. Enterprise License available for organizations with annual revenue exceeding $1M.
Not given
Key features
Stable Diffusion 3.5 Large with 8.1 billion parameters for superior quality and prompt adherence
Stable Diffusion 3.5 Large Turbo generates high-quality images in 4 steps
Model requires only 9.9 GB of VRAM to unlock full performance
Diverse outputs with different skin tones and features without extensive prompting
Generates wide range of styles including 3D, photography, painting, line art
Stable Audio 2.0 generates full tracks up to three minutes at 44.1 kHz stereo
Stable Video Diffusion transforms images into 2-second high-quality videos
Fine-tuning and customization with Query-Key Normalization for flexible base models
ControlNets for Stable Diffusion 3.5 Large including Blur, Canny, and Depth
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
What makes it different
Runs on standard consumer hardware without heavy demands
Highly customizable models designed for flexible base to build upon
Ownership of generated outputs without restrictive licensing implications
Leads market in prompt adherence and rivals larger models in image quality
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
Stability AI
Linum