gemini-3.8-flash-tts

gemini-3.8-flash-tts

Google‘s studio-quality flagship text-to-speech large model.
2026-10-14
Audio-Video Processing
Input:
$0.5/1M tokens
Output:
$9/1M tokens
Bulk order? Contact your manager for exclusive deals

API Overview

Gemini 3.8 Flash TTS is Google’s flagship text-to-speech model, designed for studio-quality natural voice, highly expressive emotional delivery, and realistic multilingual dialects. As the top-tier pillar of the Gemini API voice generation family, this model delivers exceptional vocal fidelity and natural intonation, perfectly supporting script-level multi-character dialogue, sentence-by-sentence performance control with ambient sound and emotion tags, and consistent timbre over long reading sessions. The model natively supports 130 languages and numerous regional accents, and enables rapid voice cloning from 30-second audio clips as well as text-prompted voice design, making it the ultimate cost-effective industrial-grade foundation for building high-quality AI narration, brand-specific voices, audiobooks, and game character voiceovers.

───────────────────────────────────────────────────────────────────

Core Capabilities

Studio-Level Emotional Delivery and Sentence-by-Sentence Director Controls: Achieves a complete separation of script and performance instructions. Developers can control speech rate and tone via independent style fields, and the model natively supports inserting natural vocal events such as <laugh> (laughter), <cough> (coughing), <sigh> (sighing), and <gasp> (gasping).

Long-Form Reading Timbre Stability: Completely resolves the issue of timbre drift in traditional TTS systems during long texts, maintaining consistent tone, volume, intonation, and room acoustics over several minutes or even extended narratives.

Voice Design and 30-Second Rapid Replication: Comes preloaded with 30 curated voices and a library of over 2,000 additional timbres; supports designing new voices from scratch using natural language descriptions, or quickly cloning target voices by uploading approximately 30 seconds of reference audio.

130 Languages and Dialects Covered: Supports 130 global languages, regional accents, and minority languages, with IPA international phonetic transcription coverage for correction, and can handle overlapping and interrupted dialogues among multiple characters.

API Console

Log in to explore more features! Click to Log In

API Analytics

API Reference (1)

API DescriptionAPI EndpointRequest MethodStabilityParameter Description
gemini-3.8-flash-tts
POST
Stable
View Details

API Pricing

$¥ 円 ₽
ModelDescriptionOfficial Price302.AI PriceOfficial Price Gap

gemini-3.8-flash-tts

-

Input$0.5 / 1M tokens
Output$9 / 1M tokens

Input$0.5 / 1M tokens
Output$9 / 1M tokens

Original Price