01
What is Gemini 3.8 Flash TTS?
Google announced Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS on September 23, 2026. Flash TTS is positioned for creative voice design and directed performances, while Flash-Lite targets higher-volume, cost-efficient speech generation.
Both models are rolling out through the Gemini API and Google AI Studio. Current Gemini API documentation lists the model IDs gemini-3.8-flash-tts and gemini-3.8-flash-lite-tts and distinguishes TTS from the interactive Gemini Live API.
02
Voice design, replication and performances
Flash TTS supports prompt-designed voices and a large production voice library. Google also documents line-level performance direction, non-verbal vocal events and native two-speaker scene staging for scripts such as podcasts and dialogue.
Voice replication can create a reusable vocal profile from about 30 seconds of reference audio, but Google requires the user to own the voice or have rights to use it and requires a matching verbal-consent recording from the voice owner.
03
Languages, voices and long-form use
Google's launch announcement says the family supports more than 100 languages and dialects and offers more than 2,000 production-ready voices. Exact voice availability can vary by model or product surface, so teams should verify the live API documentation for localization work.
Google positions the models for podcasts, audiobooks, games, dubbing and voice agents. Claims about long-form consistency are vendor claims rather than guarantees; production teams should test timbre, pronunciation and pacing on their own scripts.
04
SynthID, C2PA and regional restrictions
Google says audio generated by Gemini Audio models carries SynthID watermarking. Replicated voices additionally use C2PA credentials, while consent verification is built into the replication workflow.
At launch, AI Studio voice replication is unavailable in Illinois, Texas, the EEA, the United Kingdom, Switzerland and India. This restriction is specific to the documented replication workflow and should be rechecked as rollout changes.
05
API status, benchmarks and caveats
The Gemini API speech-generation guide is the current implementation authority for model IDs and request structure. Pricing, quotas and regional access are time-sensitive, so this article does not infer missing commercial details from older TTS models.
Google reports strong scores on Hume AI voice-design and accent benchmarks and says Flash TTS and Flash-Lite perform strongly in blind preference testing. Those are Google-reported launch results and should remain attributed rather than being treated as universal superiority claims.
Sources
Primary and supporting sources
Use the linked documentation to check current pricing, access and product behaviour. The dates above describe this article; they do not guarantee that every vendor claim remains current.