TENCENT's Hunyuan has launched the new-generation speech recognition model Hy ASR 3.0 preview, integrating high-precision speech recognition and deep semantic understanding capabilities. The model achieved comprehensive improvements across key dimensions including general recognition, contextual awareness and dialect coverage, enabling accurate, coherent transcription results that are closer to user intent in more complex real-world inputs. It has evolved from "word-by-word transcription and single-point optimization" to "context understanding, scenario compatibility and one-click output".

Hy ASR 3.0 preview is now available on the Tencent Cloud official website with API services open to external users. It can be widely applied in scenarios such as intelligent customer service, content understanding and voice search. Yuanbao deeply participated in the joint development and has completed the initial launch. Users can experience enhanced dialect recognition, contextual intelligent error correction and stable transcription in complex environments by pressing and holding to speak in Yuanbao. Voice input is now more accurate and stable, and is available free of charge. Products including WorkBuddy are also being integrated progressively. (jl/u)