# Gemini 3.5 Live Translate

> Low-latency audio-to-audio speech translation. Near real-time speech-to-speech across 70+ languages, preserving the speaker's intonation, pacing, and pitch. WebSocket-based realtime session.

Human version: https://kymaapi.com/models/gemini-3.5-live-translate-preview
Live JSON: `GET https://kymaapi.com/v1/models` (no auth required)

## Facts

- **Model ID**: `gemini-3.5-live-translate-preview` — pass this as `model` in the request body
- **Creator**: Google
- **Released**: 2026-06-09
- **Price**: $0.06345 / minute

## Call it

```bash
curl https://kymaapi.com/v1/audio/speech \
  -H "Authorization: Bearer $KYMA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "gemini-3.5-live-translate-preview", "prompt": "..."}'
```

## Positioning

This model provides low-latency, audio-to-audio speech translation across 70+ languages while preserving the original speaker's intonation and pacing. Reach for it when building real-time voice agents or live dubbing workflows that require WebSocket-based streaming.

## About

Gemini 3.5 Live Translate is a real-time speech-to-speech model from Google. It accepts audio input and returns translated audio output, maintaining the original speaker's pitch, pacing, and intonation across more than 70 languages.

On Kyma, it runs as a WebSocket-based realtime session. Requests automatically failover if a serving path degrades, and the platform returns exact usage costs in the response payload. Prompt caching is supported, billing repeated prefixes at 10% of the standard input rate.

It does not support text input, vision, structured outputs, or reasoning. The context window is 16,384 tokens with a maximum output of 32,768 tokens. It operates on a per-minute billing mode and sits in the premium cost tier.

## Use cases

- **Live Voice Translation** — Translates spoken audio across 70+ languages in near real-time while keeping the original voice characteristics.
- **Multilingual Voice Agents** — Powers conversational agents that listen and respond in audio without intermediate text steps.
- **Real-Time Audio Dubbing** — Generates synchronized translated audio tracks that match the source speaker's pacing and intonation.
- **Streaming Audio Conversations** — Handles continuous WebSocket sessions for interactive audio conversations with fast turnaround.

## Not ideal for

Do not use this model for text-based chat, image analysis, structured data extraction, or tasks requiring explicit reasoning steps.

## Pick something else when

- You need a lower-cost tier for high-volume audio streaming → `gemini-3.1-flash-live-preview`
- You require native audio processing with different latency characteristics → `gemini-2.5-flash-native-audio-preview-12-2025`
- You need an alternative realtime audio routing path for translation → `gpt-realtime-translate`

## See also

- All models: https://kymaapi.com/models.md
- Pricing: https://kymaapi.com/pricing.md
- Other models by Google: https://kymaapi.com/models?q=Google