Gemini 3.8 Flash Lite TTS

Overview. Gemini 3.8 Flash-Lite TTS is Google's dedicated text-to-speech model built for fast, high-throughput voice synthesis. It processes text prompts and returns standard WAV audio across 101 supported languages. Developers reach for it when powering interactive audio experiences that require quick voice generation without general conversational model overhead.

Key strengths. The model supports an 8,192-token text input window and generates up to 16,384 audio tokens, allowing long-form passage synthesis within a single call. Emitting standard RIFF-header WAV audio by default streamlines playback and media pipelines without requiring raw PCM transformations. Multilingual support across more than 100 languages ensures consistent speech output across global regions.

When to use it. Direct requests to this model through the Smart Router when your application requires interactive latency for speech synthesis rather than complex reasoning or text transformation. Use it for automated announcements, multilingual vocal prompts, and spoken interface responses where audio synthesis is the primary objective. It is ideal for high-volume audio generation tasks where function calling and code execution are not required.

Limitations. Gemini 3.8 Flash-Lite TTS is strictly an audio synthesis model and does not support text reasoning, function calling, or code execution. It accepts text inputs only, without vision or audio input capabilities, and does not support live streaming inference.

Sovereignty note. Operated by Google under US commercial terms, this proprietary model is subject to US jurisdiction and CLOUD Act data access frameworks. Workloads requiring strict Sovereign guarantees or EU data residency should assess endpoint data routing policies before transmission.