Kunlun WanweiOn April 17, 2024, the "Tiangong 3.0" base model and the "Tiangong SkyMusic" music model were announced to startPublic Beta.
Among them, "Tiangong 3.0" has 400 billion parameters, surpassing Grok-1 with 314 billion parameters, making it the world'smaximumThe open source MoE large model has achieved breakthrough performance improvements in semantic understanding, logical reasoning, versatility, generalization, uncertain knowledge, and learning ability.
At the same time, "SkyMusic" is ChinaThe firstThe music SOTA model has the ability to generate music from end to end, including instruments, vocals, and melodies. It performs well and supports multiple dialects, surpassing overseas models.top notchThe AI music model Suno V3 has become the world's AI music SOTA model.
"Tiangong 3.0" has AI capabilities such as multi-round search and comprehensive tool calls, chart drawing, research mode, enhancement mode, image modification and expansion, providing a more powerful AI experience. At the same time, "Tiangong SkyMusic" supports reference music generation and dialect song generation, allowing users to easily create music.
Kunlun Wanwei continues to launch cutting-edge large-scale model products and build an AI business matrix, including AI search, AI music, AI video, etc., becoming a leader in technology and engineering capabilities among domestic AI technology companies.Strongest, one of the most comprehensive layouts.