STORM-Ensemble 3.0 is a speech synthesis model that generates natural speech in 64 African languages using 166 distinct voices produced entirely by the system.