Tencent Hunyuan Releases Hy ASR 3.0 Preview Speech Recognition Model with ~3% Multilingual Word Error Rate
Tencent
Tencent Hunyuan has unveiled Hy ASR 3.0 preview, a speech recognition model achieving approximately 3% word error rate across multiple languages. The announcement was reported by PChome, a Chinese tech media outlet.
Tencent Hunyuan, the AI division of Tencent, has introduced the Hy ASR 3.0 preview, a new speech recognition model. This model reportedly achieves a word error rate of about 3% across multiple languages, indicating high accuracy in transcribing speech. The news was published by PChome, a Chinese technology news website. Further technical details or release timelines were not provided in the original report.
Source: Tencent Hunyuan (GNews) —
original