Fish Audio Launches Drama 3, an Open-Weight Text-to-Speech Model with Performance Directions

Fish Audio released Drama 3 on October 13, 2026—an open-weight text-to-speech model that allows users to direct vocal performance through embedded natural-language instructions. The system uses a 35-billion-parameter mixture-of-experts architecture that activates approximately 3 billion parameters per token, initialized from Qwen3.5-35B-A3B, and operates at 271 frames per second on a single H200 GPU, more than seven times faster than Fish Audio's S2 predecessor despite significantly larger parameter count.