Google DeepMind has introduced new voice synthesis tools capable of generating highly expressive and scalable custom audio. The released Gemini 3.8 Flash and Flash-Lite TTS models allow creators and developers to design unique voices with specific accents and support voice replication across 100 languages via AI Studio and the Gemini API. Gemini 3.8 Flash TTS leads the Artificial Analysis Pronunciation Robustness benchmark with a score of 89.5% and holds the No. 2 position on the Provider Voice Arena Leaderboard with an Elo of 1,263.

Pricing for Gemini 3.8 Flash TTS is $32.98 per 1M characters, while the Flash-Lite version is $22.07 per 1M characters, making both cheaper than Eleven v3 at $100 per 1M characters. The models include a library of over 2,000 production ready voices and process up to 44.1 characters per second. Despite these gains, Google's models remain slower than rivals like Falcon 2, which processes 204.9 characters per second.