Stability AI unveils audio model capable of generating six-minute songs
Stability AI, the creator of Stable Diffusion, has unveiled a new series of audio models named Stability Audio 3.0. According to the company, the flagship model is capable of producing professional-quality music tracks exceeding six minutes in length.
Under the Stability Audio 3.0 umbrella, the company is introducing four models: small SFX (459M parameters), small (459M parameters), medium (1.4B parameters), and large (2.7B parameters). The two small models are designed for on-device sound and music generation, with a maximum output length of two minutes.
The medium and large models can produce full compositions up to 6 minutes and 20 seconds, preserving musical structure and melodic coherence. That's more than twice the duration achievable by Stable Audio 2.0, which launched in 2024.
Stability AI is releasing the small SFX, small, and medium models with open weights, allowing anyone to use and modify them. In 2024, the company introduced Stable Audio Open, which enabled music generation up to 47 seconds. This new model family represents a significant advancement over its open-source predecessors.

Image credits: Stability AIImage credits: Stability AI
The large model is accessible only via the API and paid self-hosting services. Additionally, businesses with annual revenue exceeding $1 million are required to obtain an enterprise license.
Numerous companies, such as Google and ElevenLabs, are launching models and tools for music generation. However, as the ongoing legal disputes involving Suno and Udio have demonstrated, data licensing and partnerships with music labels may be crucial for the long-term viability of these services.
Last year, Stability AI signed agreements with Warner Music Group and Universal Music Group to develop models and music creation tools. The company states that its latest audio models are trained on fully licensed data.
The AI startup is developing a new suite of products tailored for professional musicians, though it has not disclosed specific features. Ethan Kaplan, previously chief digital officer at Universal Audio and Fender, is joining the company to head Stability's professional music division.
Several AI companies are strengthening their credibility by recruiting music industry executives. Earlier this year, Suno appointed former Merlin CEO Jeremy Sirota as chief commercial officer. ElevenLabs also brought on Derek Cournoyer from indie music publisher Kobalt as a strategy lead for its music business.
Related article
Spotify Touts AI as Key to Empowering Its Leading Developers
Has AI-driven development reached a critical milestone? Spotify certainly suggests so. During its Q4 earnings call this week, the company revealed that its top engineers "haven’t written a single line of code since December." That remark came from Sp
Janet Jackson's 'Rhythm Nation' caused select Windows laptops to crash for years
Longtime readers of The Verge may remember the peculiar incident in which Janet Jackson's "Rhythm Nation" music video could crash certain Windows laptops simply by being played nearby. Now, in a blog post highlighted by PCWorld, Microsoft employee Ra
SoundCloud Clarifies It Does Not Train AI on User Music
In February of last year, the music-sharing platform SoundCloud discreetly revised its terms of use, introducing new provisions that permit the training of AI models using user‑generated material, as reported by TechCrunch. Although the company state
Related Special Topic Recommendations
Comments (0)
0/500
Stability AI, the creator of Stable Diffusion, has unveiled a new series of audio models named Stability Audio 3.0. According to the company, the flagship model is capable of producing professional-quality music tracks exceeding six minutes in length.
Under the Stability Audio 3.0 umbrella, the company is introducing four models: small SFX (459M parameters), small (459M parameters), medium (1.4B parameters), and large (2.7B parameters). The two small models are designed for on-device sound and music generation, with a maximum output length of two minutes.
The medium and large models can produce full compositions up to 6 minutes and 20 seconds, preserving musical structure and melodic coherence. That's more than twice the duration achievable by Stable Audio 2.0, which launched in 2024.
Stability AI is releasing the small SFX, small, and medium models with open weights, allowing anyone to use and modify them. In 2024, the company introduced Stable Audio Open, which enabled music generation up to 47 seconds. This new model family represents a significant advancement over its open-source predecessors.

Image credits: Stability AIImage credits: Stability AI
The large model is accessible only via the API and paid self-hosting services. Additionally, businesses with annual revenue exceeding $1 million are required to obtain an enterprise license.
Numerous companies, such as Google and ElevenLabs, are launching models and tools for music generation. However, as the ongoing legal disputes involving Suno and Udio have demonstrated, data licensing and partnerships with music labels may be crucial for the long-term viability of these services.
Last year, Stability AI signed agreements with Warner Music Group and Universal Music Group to develop models and music creation tools. The company states that its latest audio models are trained on fully licensed data.
The AI startup is developing a new suite of products tailored for professional musicians, though it has not disclosed specific features. Ethan Kaplan, previously chief digital officer at Universal Audio and Fender, is joining the company to head Stability's professional music division.
Several AI companies are strengthening their credibility by recruiting music industry executives. Earlier this year, Suno appointed former Merlin CEO Jeremy Sirota as chief commercial officer. ElevenLabs also brought on Derek Cournoyer from indie music publisher Kobalt as a strategy lead for its music business.
Spotify Touts AI as Key to Empowering Its Leading Developers
Has AI-driven development reached a critical milestone? Spotify certainly suggests so. During its Q4 earnings call this week, the company revealed that its top engineers "haven’t written a single line of code since December." That remark came from Sp
Janet Jackson's 'Rhythm Nation' caused select Windows laptops to crash for years
Longtime readers of The Verge may remember the peculiar incident in which Janet Jackson's "Rhythm Nation" music video could crash certain Windows laptops simply by being played nearby. Now, in a blog post highlighted by PCWorld, Microsoft employee Ra





Home






