QueueMix release voice recognition model Hy ASR 3.0 preview

On 4 August, a new generation of voice recognition models, Hy ASR 3.0 preview, was officially released by Quetzuan. Based on the language understanding capability of the latest-generation large-language model Hy3, the model is able to provide accurate, consistent and closer-to-the-user intent transliteration results in more complex real input. The model is now online and provides API services to the outside world, which can be widely applied to intelligent passenger service, content understanding and voice search. Wonbao has already started off online and is free of charge; products such as WorldBuddy are also being accessed。