Stable Audio Open 1.0
Stable Audio Open 1.0 converts English text prompts into variable-length stereo audio. It combines a waveform autoencoder, T5 text conditioning and a diffusion transformer operating on audio latents.
Model design and access
The card lists 486,492 recordings from Freesound and the Free Music Archive. It supports the stable-audio-tools and Diffusers libraries. This earlier open-weight release uses a different toolchain and text encoder from Stable Audio 3.Stable Audio Open 1.0 official model documentation
Practical use
Practical guidance. Open 1.0 is useful for sound design experiments and existing local pipelines that need short audio. The 47-second cap calls for planned segmentation in longer work. Listen for unwanted speech-like sounds and abrupt endings, and trim effects to their actual use rather than treating the requested duration as an editing boundary.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Model identifiers | stable-audio-open-1.0 |
| Output | Up to 47 seconds at 44.1 kHz stereo.Stable Audio Open 1.0 official model documentation |
| License | Stability AI Community License.Stable Audio Open 1.0 official model documentation |
Related models
- Stable Audio 3.0 Small SFX
Stability AI · Stable Audio
A compact text-to-audio model for sound effects with downloadable weights.
Launch date not confirmedDownloadable - Stable Audio Open Small
Stability AI · Stable Audio
Stable Audio Open Small is an earlier compact text-to-audio model for short clips.
Launch date not confirmedDownloadable