Google - Kyma API

4 Google audio models on Kyma. See Audio for the endpoint, parameters and billing.

Model ID	Best for	Per minute
`gemini-3.5-live-translate-preview`	Live speech-to-speech translation, multilingual voice agents, dubbing	$0.06
`gemini-3.1-flash-live-preview`	Voice agents migrating off Gemini 2.5 native audio, low-latency spoken dialogue	$0.04
`gemini-3-flash-audio`	Audio scene understanding, mood/emotion detection, music recognition	$0.0026
`gemini-2.5-flash-native-audio-preview-12-2025`	Conversational voice agents, multilingual chat, voice picker UX	$0.04

Retiring

gemini-2.5-flash-native-audio-preview-12-2025 is served until 2026-10-20. Move to another model before then.