Stability AI, the company best known for its image model Stable Diffusion, is now making waves in the mobile audio space. It has launched Stable Audio Open Small, a lightweight AI model that can generate stereo audio directly on smartphones—without relying on the cloud.
This release comes as part of a collaboration with Arm, the chip designer whose processors power billions of mobile devices. Unlike popular audio-generation tools like Suno and Udio, which require internet connectivity due to their reliance on cloud-based processing, Stability’s new model runs locally. That means users can create audio clips even when offline.
Built specifically for low-latency performance, Stable Audio Open Small clocks in at 341 million parameters and is optimized for Arm CPUs. According to Stability, it can generate up to 11 seconds of stereo sound—like drum loops or short instrumental riffs—in under 8 seconds on a smartphone.
What truly sets this model apart is its dataset. Stability trained it exclusively on royalty-free content from the Free Music Archive and Freesound, avoiding copyrighted material altogether. That’s a significant contrast to its competitors, who have faced scrutiny over using copyrighted music in training—potentially opening them up to legal challenges.
Despite its speed and portability, the model comes with a few caveats. For one, prompts must be written in English. It also doesn’t yet support vocals or the ability to create full-length, studio-quality tracks. In fact, the model performs best with short audio effects and struggles with musical genres outside of its Western-leaning training set.
Another limitation? Licensing. While individuals, hobbyists, and small businesses (those earning under $1 million annually) can use Stable Audio Open Small for free, larger developers and companies will need to pay for an enterprise license. That could limit adoption for more commercial use cases.
This launch comes at a pivotal time for Stability AI. After internal turmoil and financial missteps under former CEO Emad Mostaque, the company faced setbacks including failed partnerships and a flurry of staff departures. But since appointing a new CEO, Stability has taken steps to rebuild. Its latest moves include adding filmmaker James Cameron to its board and releasing new image-generation models alongside this audio innovation.
With Stable Audio Open Small, Stability is positioning itself not just as a leader in generative imagery, but as a serious contender in AI-powered sound—especially on mobile.
