Gemini 3.1 Flash TTS

gemini-3.1-flash-tts is a text-to-speech (TTS) model released by Google, part of the Gemini 3.1 Flash series, with audio as its output modality. Google positions it as a low-latency, highly natural speech generation model, with a focus on speech controllability and multilingual coverage.

According to official information, the model supports fine-grained control over voice style, speaking rate, accent, and delivery through natural-language "audio tags" embedded in the text, covers more than 70 languages, and outputs high-fidelity speech; the generated audio embeds a SynthID watermark to identify AI-generated content. Officially listed use cases include accessibility and assistive communication, game and audio content production, multilingual customer service and announcements, and localized dubbing—speech scenarios that require controllable delivery.

The model is currently in public preview; the official version is offered as gemini-3.1-flash-tts-preview on Google AI Studio and Vertex AI, and its interface and behavior may still change. On HiAPI.ai, the model is not yet available for calls and is in a coming-soon state; the specific launch date, billing method, and API details are subject to the official release.

Once launched, HiAPI.ai will provide this speech synthesis capability through the platform's unified model integration approach; the actually available parameters, language coverage, and invocation forms are subject to the official general release and the platform's launch documentation.