Riding the surging wave of the intelligent agent era, UNISOUND (09678) is rapidly completing the capability puzzle of its "One Base, Two Wings" technology strategy. Following the upgrade of its "Shanhai Zhiyi" 5.0 medical large model earlier this year, the company has now重磅 launched the "Shanhai Zhiyin" large model 2.0.

It is reported that the "Shanhai Zhiyin" large model 2.0 leverages the multimodal and cross-lingual foundational capabilities of "Shanhai Atlas," with a core focus on three major evolutionary capabilities: understanding professional terms and local accents, conversing with familial warmth and empathy, and delivering exceptionally quick-witted responses. In the dimension of "understanding," the model's ASR capability has demonstrated leading speech recognition performance in both public test sets and proprietary full-scenario test sets, achieving a leading level from general to extreme comprehensiveness in evaluations, surpassing mainstream domestic open-source and closed-source speech large models and reaching the industry's highest standard.

On the "expression" front, Shanhai Zhiyin-TTS, centered on "highly human-like + creatively diverse" principles, currently supports 12 dialects (including Cantonese, Sichuanese, and Shanghainese) and 10 foreign languages, and can even switch between 12 different Mandarin speaking styles. More importantly, based on an end-to-end interaction brain, Shanhai Zhiyin 2.0 has overcome the challenge of smooth full-duplex interaction, supporting随时 interruption, immediate response, and coherent follow-up questions, making human-machine dialogue as fluid as a conversation between close friends.

Underpinning all these capabilities is UNISOUND's独创 "Shanhai Atlas" intelligent computing integrated base, which deeply integrates a general multimodal large model foundation with the Atlas infrastructure. This serves as the foundation for professional intelligent agents and is also the core of perceptual AI.