Google DeepMind released its newest text-to-speech models, Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, for developers and enterprises via the Gemini API and AI Studio. The Flash TTS model hit the No. 1 spot on the Artificial Analysis Pronunciation Robustness Benchmark with a score of 89.5% and ranked No. 2 on the Provider Voice Arena Leaderboard with an Elo of 1,263. These tools support voice replication and more than 2,000 production-ready voices, allowing users to generate multi-voice conversations with controls for tone, pacing, and natural cues like laughter.

The two models differ in targeting and cost, with Flash priced at $32.98 per 1 million characters for creative production and Flash-Lite at $22.07 per 1 million characters for scale and voice agents. While both are more expensive than the previous Gemini 3.1 Flash TTS at $18.31 per 1 million characters, they remain cheaper than Eleven v3 at $100 per 1 million characters. In terms of performance, Flash processes 44.1 characters per second and Flash-Lite processes 40.2 characters per second, both exceeding realtime speeds.