LiveKit says Gemma 4 31B hits 192 ms to first token in voice agents
GlennCameronjr · x · 2026-07-30
LiveKit says the fastest LLM for a voice agent right now is an open-weight model: Gemma 4 31B on LiveKit Inference reaches first token in 192 ms and a full spoken sentence in 354 ms, while costing $0.40 per 1M input tokens ($0.20 cached) and $1.20 per 1M output tokens.
The post frames this as a practical latency-plus-cost win for real-time voice agents, not just a model-quality comparison.
More from Infra
- NVIDIA launches RTX Spark for local AI on Windows PCs — ostrisai · 2026-10-16
- SkyRL ships Day-0 RL support for Beam with +62% LoRA decode throughput in vLLM — vllm_project · 2026-10-16
- DFlash2 delivers up to 3.03x decode-only speedup for Beam across five benchmarks — vllm_project · 2026-10-16
- vLLM's Beam serving kernels cut GB200 forward time from 8.15 to 6.47 ms at TP8 batch 128 — vllm_project · 2026-10-16
- Reflection's 501B MoE Beam hits 463 tok/s on 8x GB200 with NVFP4 at day 0 — zhyncs42 · 2026-10-16
- Nemotron 3.5 Bolt runs 16 concurrent streams on a single DGX Spark — NVIDIAAI · 2026-10-16