Stability AI has introduced a new version of the audio generator — Stable Audio 2.5. This model is designed for companies that aim to create their own “sound identity.” Stable Audio 2.5 allows for the quick generation of custom music tracks up to three minutes long. Users can set sound parameters using natural language, such as specifying an “uplifting” mood for the track.
The model can create complex compositions with an introduction, main part, and conclusion. Additionally, a new “inpainting” feature allows users to upload their own audio fragment and let the AI enhance it with new elements. The system automatically rejects attempts to use copyrighted materials.
Stable Audio 2.5 was trained on a fully licensed database. This allows the created tracks to be used in advertising, retail spaces, or other business projects without the risk of copyright infringement. The company offers model customization for a specific brand sound library so that music and sound effects are easily recognizable as part of the brand’s style.
Users can try Stable Audio 2.5 with options: free with a limit of up to 10 tracks per month, Pro plan for $12 with a limit of 250 tracks, or use the extended Studio and Max packages. The model is aimed at companies that value the speed of creating high-quality sound content for their own needs.
