Stability AI Revolutionizes Music with Stability Audio 3.0
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
Stability AI, known for its Stable Diffusion artificial intelligence model, has recently launched Stability Audio 3.0, a new range of audio models capable of generating professional-quality music tracks with an impressive duration of over six minutes.
A New Family of Audio Models
Stability AI's new offering includes four distinct models under the name Stable Audio 3.0:
- small SFX (459 million parameters)
- small (459 million parameters)
- medium (1.4 billion parameters)
- large (2.7 billion parameters)
The small models are designed to generate sounds and music directly on devices for durations of up to two minutes. In contrast, the medium and large models can create complete compositions of 6 minutes and 20 seconds, more than double the duration that could be produced by the previous version, Stable Audio 2.0, launched in 2024.
Accessibility and Usage Conditions
Stability AI has chosen to make the small SFX, small, and medium models available with open weights, allowing anyone to use and modify them. However, the large model is reserved for use via an API and paid hosting services. Companies generating over $1 million in revenue will also need to obtain a business license to use it.
In 2024, the company launched Stable Audio Open, which allowed music generation up to 47 seconds. This new family of models thus represents a significant advancement over previous versions.
Partnerships and Legal Issues
The company has already established partnerships with music giants such as Warner Music Group and Universal Music Group to develop its models and music creation tools. Stability AI assures that its latest set of audio models is based on fully licensed data, a crucial point as licensing issues and partnerships with music labels become essential for the sustainability of these services.
Strengthening Skills in the Music Industry
To support its expansion into AI-generated music, Stability AI has recruited Ethan Kaplan, former digital director at Universal Audio and Fender, to lead its professional music offering. This move is part of a broader trend where AI companies are seeking to strengthen their expertise by hiring leaders from the music industry. Suno and ElevenLabs have also followed this path by hiring Jeremy Sirota and Derek Cournoyer, respectively, for strategic roles.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.