News Summary

Stability AI, in collaboration with Arm, has launched Stable Audio Open Small, a text-to-audio generative AI model designed to run efficiently on smartphones. This breakthrough allows audio generation directly on mobile devices without requiring internet connectivity, thanks to its optimization for Arm CPUs. The model, consisting of 341 million parameters, can produce up to 11 seconds of audio in under 8 seconds on a smartphone, making it significantly faster than previous models. Stability AI’s partnership with Arm, which powers the majority of smartphones globally, utilizes Arm’s software stack to enhance performance, achieving a 30x faster response time. Despite its speed and efficiency, the model has some limitations, including restrictive usage terms for companies with high revenue and a bias towards Western musical styles due to its training data. Stability AI, known for its image-generation model Stable Diffusion, has faced financial challenges but continues to innovate under new leadership.