Gemini 3.8 text-to-speech says hello
deepmind ·
Google DeepMind has introduced Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, marking a significant advancement in expressive audio generation within the Gemini family. These models allow users to create custom character voices from scratch or replicate existing ones using simple natural language prompts across more than 100 languages. The Flash version is designed for deep creative direction, offering granular control over pacing, emotion, and realistic conversational sounds like laughter or sighs. Meanwhile, the Flash-Lite variant focuses on high-volume, cost-efficient scaling for tasks such as dubbing and voice agents. Both models are now available in Google AI Studio and the Gemini API, with enterprise access coming soon via Gemini Enterprise. To ensure responsible use, the system includes built-in safety features such as consent verification for voice replication and SynthID watermarking for all generated audio. Early benchmarks indicate that these models lead in overall quality and accent modeling compared to previous iterations.
deepmind