BREAKING
LiveKit Tunes Gemma 4 31B for Voice
0
ms
Time to first token
0
ms
Time to first audio
Tool Use vs Task Completion
tau2bench
76.9
Hotel tasks done
88
Drop-In Voice Pipeline
1
Deepgram STT
↓
2
Gemma 4 31B
↓
3
Cartesia TTS
Hosted or Self-Host
LiveKit Inference
Hosted
●
$1.20 per million tokens
●
Python and TypeScript SDKs
Self-Host
Open-weight
●
Apache 2.0 license
●
On Hugging Face and AWS Bedrock
Speed Meets Tool Use
AI NEWS BLITZ
LiveKit has launched a latency-optimized Gemma 4 31B for real-time voice AI agents.