Japanese Startup higurashi Open-Sources Full-Duplex Voice Model Haru 4B
higurashi releases Haru 4B, a 4B-parameter full-duplex speech model under Apache 2.0, aiming to solve latency and non-verbal cue loss in traditional cascaded voice systems.
Japanese AI startup higurashi released Haru 4B, a full-duplex speech dialogue model, on November 13, 2026. The company has made the pre-trained weights and inference code available on Hugging Face under the Apache 2.0 license. With 4 billion parameters, the model is specifically designed to address the needs for naturalness and low latency in real-time voice interactions.
Unlike traditional "cascaded" architectures that sequentially process Automatic Speech Recognition (ASR), Large Language Models (LLM), and Text-to-Speech (TTS), Haru 4B employs an end-to-end full-duplex LSLM (Large Speech Language Model) technology. This architecture allows the model to simultaneously process three data streams—user voice, model voice, and text—synchronizing information approximately every 133 milliseconds. This enables the model to "listen while speaking," facilitating low-latency responses to user interruptions, backchanneling, and topic shifts, while preserving paralinguistic features such as tone and emotion.
According to higurashi's official announcement, Haru 4B was pre-trained on 40 million hours of multilingual data including Japanese and English, currently supporting nine languages. The model also incorporates the VibeVoice continuous speech tokenizer to balance audio quality with computational efficiency. Despite its fluent conversational capabilities, higurashi explicitly notes that due to its 4B parameter size, the model's logical reasoning ability is roughly equivalent to GPT-2 levels, highlighting that its primary value lies in its interaction mechanism rather than complex task processing.