AI product
A voice-tuned version of the Gemma 4 31B language model hosted on LiveKit Inference for production voice agents. The talk describes it as designed for long prompts, many tool calls, and tight latency budgets; it reports an 88% voice-evaluation score and speech beginning in about 380 milliseconds.
1 use taken from transcripts — each links to the moment in the video.
A voice-tuned version of the Gemma 4 31B model hosted on LiveKit Inference, tuned for production voice agents with long prompts, many tools, and tight latency budgets. Passes voice evals at 88% and starts speaking in about 380 milliseconds — the fastest high-quality voice LLM available, per the talk.
1 in the library.