Google releases Gemini 3.8 Flash TTS models with voice replication

Google has released two new production-ready text-to-speech (TTS) models: Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS. These models are part of the Gemini API and Google AI Studio ecosystem.

The update introduces a comprehensive 'Voices' layer, which includes custom voice personas, the ability to replicate voices with user consent, and a library of over 150 ready-made and custom voices. The Gemini 3.8 Flash model is designed for high expression, multi-speaker dialogues, and long-form narration, while the Flash-Lite version is optimized for speed and high-volume generation.

Both models support multiple languages, including Polish. Promotional pricing for the API is available until December 31, 2026, after which standard rates are expected to double.