Fun-ASR-Realtime (Tongyi Qianwen Speech)

Overview

Alibaba’s Qianwen ecosystem recently released **Fun‑ASR‑Realtime**, a real‑time automatic speech recognition model optimized for Chinese and bilingual call‑center, meeting and live‑streaming scenarios.[8] It is designed for low latency transcription and integrates with Alibaba Cloud’s DashScope platform, enabling developers to build voice bots, live captions and searchable audio archives.[8] The model complements Tongyi Qianwen LLMs to power full voice‑to‑intent pipelines for enterprises.