Description

Haru-4B is an open-source voice model designed for full-duplex communication, allowing it to listen and talk simultaneously like a human phone call. It effectively manages natural conversation flow including backchannels, interruptions, and turn-taking, achieving approximately 65ms latency on consumer hardware like a MacBook.

Descriptions, tags, and model credits may be AI-generated or inferred from public sources and can be incomplete or wrong. Learn more.