AI Development
Google launches Gemini 3.8 Flash TTS and Flash-Lite TTS models
Editorial Analysis
Google released two new text-to-speech models supporting promptable voice design from natural language descriptions, over 2,000 production voices, 30-second consented voice replication with SynthID and C2PA provenance, and support for 100-plus languages. Flash ranked first on independent voice design benchmarks; the models are available via Gemini API and AI Studio for creative and high-volume audio generation use cases.
At a Glance
Date
September 23, 2026
Importance
Medium
3/5
Category
model
Axis of Change
Capability Gain
Organizations
Models Affected
Sources
- 01 blog.google
- 02 ai.google.dev