#TTS
2 published articles
Drama 3 puts directed TTS into open weights under a revenue gate
Fish Audio shipped a 35B-A3B speech model you brief in brackets and scope to single words. Weights are on Hugging Face. The Community License and invite-only editing still decide who gets the full pitch.
Google ships Gemini 3.8 Flash and Flash-Lite TTS on 23 September for API and AI Studio
Generative voice design, 2,000+ library voices, and consented 30-second replication. Voice cloning blocked in several regions.