Google Unveils Gemini 3.8 Flash TTS and Flash-Lite TTS, Achieving Top Spot on Voice Design Benchmark

Google has expanded the Gemini Audio lineup by releasing Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS. Both models ranked highly on Hume AI's Voice Design Benchmark, with Flash TTS specifically recording a top score of 71.4.

Key Features and Performance

• Generative Voice Design: Users can customize roles, intonation, and vocal characteristics across more than 100 languages and dialects using natural language prompts.
• Voice Cloning: Replicates consistent vocal profiles from 30-second audio samples, protected by consent verification and SynthID watermarking.
• Performance Control: Adjusts line-by-line delivery through dialogue instructions or natural script cues, minimizing speaker drift even in long-duration audio.
• Multi-Speaker Support: Enables scenes where two speakers converse naturally within a single script.

Deployment and Partnerships

Developers can access the models immediately via the Gemini API and Google AI Studio, while enterprises will receive access through the Gemini Enterprise API later. General users can access them via Gemini Notebook (Flash TTS) and Google Vids (Flash-Lite TTS). Major partners such as Figma, HeyGen, and Linguana are integrating these models to build global dubbing and interactive agents.

Limitations

The voice cloning feature via AI Studio is not available in Illinois, Texas, the EEA, the UK, Switzerland, or India.