Jul 17, 2026
•
7 min read
354ms to First Audio: Gemma 4 31B on LiveKit Changes the Math for Voice AI
LiveKit has launched Gemma 4 31B on its Inference platform, delivering 354ms time to first audio and 192ms time to first token — while beating GPT-4.1 on agentic tool-use benchmarks. For developers building real-time voice agents, the numbers are hard to ignore.
Ab
Abhinav Ramaswamy