Stability AI, CompVis, and Runway publicly released Stable Diffusion v1 under the CreativeML OpenRAIL-M license. Developers could run and adapt the weights on their own hardware subject to the license terms.
Stability AI
Also known as: Stability · Stable Audio · Stable Audio Open · Stable Virtual Camera
Plain English
The generative AI company that co-released Stable Diffusion and helped spark the open-weight visual AI ecosystem. Today it develops image, video, 3D, and audio foundation models for creators, edge devices, and studio production pipelines.
Technical Overview
Founded in 2019, Stability AI funded and distributed Stable Diffusion weights on Hugging Face, catalyzing local Safetensors Checkpoints, LoRA adapters, and node workflows in ComfyUI. Across 2023–2026, the company expanded beyond U-Net and MMDiT image backbones into timing-conditioned latent diffusion for music and sound (Stable Audio, Stable Audio Open, Stable Audio 3.0) and multi-view 3D/video synthesis (Stable Virtual Camera), distributing weights under the tiered Stability AI Community License.
Timeline (6)
2022
Release2023
ReleaseStability AI launched Stable Audio, which uses timing-conditioned latent diffusion to generate variable-length stereo music and sound effects from text.
2024
MilestoneStability AI announced new funding, appointed Prem Akkaraju as CEO, and named Sean Parker executive chairman. The company said the recapitalization would support further development and commercialization of its generative models.
2025
ReleaseStability AI and Arm released Stable Audio Open Small, a 341-million-parameter model for short audio samples. The release paired a smaller model with Arm optimizations aimed at running it on mobile CPUs.
2026
ReleaseStability AI introduced Stable Audio 3.0 with downloadable and hosted variants for longer instrumental music and sound effects. The smaller checkpoints allow local use, while the larger hosted variant targets longer compositions.
- Milestone
Stability AI closed a $76M Series B round (bringing total funding under CEO Prem Akkaraju to $232M) backed by Electronic Arts, Sony Music Group, Universal Music Group, Warner Music Group, AMD Ventures, and Coatue to develop rights-cleared multimodal models and edge NPU deployments for music, gaming, and film production.
See also (14)
Open full graphStability AI
9 more connections
- Video Generation Models (Sora, Veo, Kling, Wan)Incoming · Uses
- CheckpointsOutgoing · Mentions
- ComfyUIOutgoing · Mentions
- Diffusion Transformer (DiT)Outgoing · Mentions
- Hugging FaceOutgoing · Mentions
- LoRA & QLoRAOutgoing · Mentions
- Music Generation ModelsOutgoing · MentionsIncoming · Mentions
- RunwayOutgoing · Mentions
- SafetensorsOutgoing · Mentions