Stability AI Launches Smartphone-Ready Audio Generator
Stability AI has released Stable Audio Open Small, a new AI model capable of generating audio directly on smartphones. This "stereo" audio-generating model is touted as the fastest on the market, efficient enough to run on mobile devices without cloud processing.
Partnership with Arm and Royalty-Free Data
Developed in collaboration with chipmaker Arm, Stable Audio Open Small addresses the offline audio generation gap. Unlike cloud-based alternatives like Suno and Udio, this model operates locally. Importantly, its training data comes entirely from royalty-free libraries like Free Music Archive and Freesound, mitigating copyright concerns.
Speed and Performance
The 341 million parameter model is optimized for Arm CPUs. It specializes in generating short audio samples and sound effects, such as drum and instrument riffs. Stability AI claims it can produce up to 11 seconds of audio on a smartphone in under 8 seconds.
Audio Samples
Listen to examples generated by Stable Audio Open Small:
Limitations and Usage Terms
Currently, Stable Audio Open Small only supports English prompts. Stability AI acknowledges limitations in generating realistic vocals or high-quality songs. The model's performance also varies across musical styles due to its Western-biased training data. While free for researchers, hobbyists, and businesses with under $1 million in annual revenue, larger organizations require a paid enterprise license.
Stability AI's Continued Development
Following financial challenges and leadership changes, Stability AI continues to innovate. The release of Stable Audio Open Small follows new image generation models and strategic appointments, signaling a renewed focus on product development.
