Breaking

Google Unveils Gemini 3.5 Live Translate Voice Translation Model

June 9, 2026 at 11:17 EDT

Foundation Models
Media Generation
AI Agents

Google DeepMind on June 9, 2026 announced its latest voice translation model, "Gemini 3.5 Live Translate." The model brings near real-time, natural speech translation to Google AI Studio, the Google Translate app and Google Meet.

June 9, 2026 · Google

Gemini 3.5 Live Translate: Near Real-Time Speech-to-Speech

Google begins beta rollout of voice translation that converts two-way conversations end-to-end — preserving the speaker's tone, pacing and emotional nuance — across Translate, Meet and AI Studio.

70+

languages supported

2,000+

language pairs

surfaces: Translate · Meet · AI Studio

API pricing — input vs output

Per equivalent unit. Output costs 6× the input.

$3.50

Input

$21.00

Output

How the audio flows

A single native-audio Gemini 3.5 model — no visible text step.

Speaker voice in

→

Gemini 3.5 audio-to-audio model

→

Translated voice out — tone preserved

Praised for

Fluent, natural-sounding speech
Instant multilingual travel chat
Live international meetings

Still rough

Turn-by-turn, not fully simultaneous
Some translation inaccuracies in Meet
Region/platform limits, variable latency

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.