Gemini 3.8 Flash-Lite TTS

Gemini 3.8 Flash-Lite TTS is the lower-cost half of Google’s Gemini 3.8 text-to-speech pair, announced on 23 September 2026 alongside Gemini 3.8 Flash TTS and listed as generally available in the Gemini API changelog on 22 September. It is aimed at high-volume work — dubbing, audio content and real-time voice agents that chain speech recognition, a language model and speech synthesis. The model page lists 101 languages with automatic detection, the same 8,192-token input and 16,384-token output limits as the larger model, and batch, caching and flex or priority inference. Google names it as the replacement for the Gemini 3.1 Flash TTS preview and says it shares a request schema with Gemini 3.8 Flash TTS, so moving between the two is a change of model name; no shutdown date for the preview was published when this was read. Voice replication is offered with the same consent check and the same regional exclusions, including the UK and the EEA. On the paid tier it costs $0.50 per million input tokens, $6.00 per million audio output tokens and $0.125 per million cached input tokens through 31 December 2026, and each rate doubles on 1 January 2027 — so a budget built on the launch price is a budget for three months. Artificial Analysis’s independent Provider Voice Arena ranked it sixth, with an Elo rating of 1,235, when read on 24 September 2026.