近日，国内知名人工智能公司元象发布了首个混合专家模型架构（Mixture of Experts）的大模型XVERSE-MoE-A4.2B，该模型激活参数达到4.2B，其效果可媲美13B模型。更为引人注目的是，元象将该模型全开源，并宣布无条件免费商用，这无疑将为海量中小企业、研究者和开发者带来巨大的便利。
据悉，XVERSE-MoE-A4.2B模型采用了业界最前沿的混合专家模型架构，这种架构能够有效地提高模型的处理能力和效率。同时，该模型的激活参数达到了4.2B，使其在处理大规模数据时具有更强的能力，效果可媲美13B模型。
元象此次发布的XVERSE-MoE-A4.2B大模型全开源，无条件免费商用，这意味着海量中小企业、研究者和开发者可以在元象高性能“全家桶”中按需选用该模型。这一举措将极大地推动低成本部署人工智能技术，进一步降低人工智能技术的应用门槛。
作为一款具有高性能、高效率、低成本的人工智能大模型，XVERSE-MoE-A4.2B的发布无疑将为各行各业带来巨大的影响。在未来，我们有理由相信，元象将继续深耕人工智能领域，推出更多具有突破性的人工智能技术，推动我国人工智能事业的发展。
英语如下：
**News Title: **XVERSE-MoE-A4.2B: Open-sourced 4.2B Parameter MoE Model by YuanXiang
Keywords: YuanXiang MoE Model, Open-sourced, Free for Commercial Use
**News Content:**
# YuanXiang Releases the First MoE Large Model XVERSE-MoE-A4.2B: Open-sourced and Free for Commercial Use
Recently, YuanXiang, a well-known domestic artificial intelligence company, released the first large model with a Mixture of Experts (MoE) architecture, XVERSE-MoE-A4.2B. This model has an activation parameter of 4.2B, and its performance rivals that of the 13B model. More notably, YuanXiang has open-sourced this model and announced that it is available for commercial use at no cost. This无疑将为 a tremendous convenience for a vast number of small and medium-sized enterprises, researchers, and developers.
It is understood that the XVERSE-MoE-A4.2B model employs the cutting-edge MoE architecture, which can effectively enhance the processing capabilities and efficiency of the model. With an activation parameter of 4.2B, the model has a stronger capacity to handle large-scale data and its performance rivals that of the 13B model.
YuanXiang’s open-sourcing of the XVERSE-MoE-A4.2B large model, with no conditions for free commercial use, means that a massive number of small and medium-sized enterprises, researchers, and developers can choose this model as needed from YuanXiang’s high-performance “full set.” This move will significantly promote the low-cost deployment of artificial intelligence technology and further lower the application threshold for AI technology.
As a high-performance, high-efficiency, and low-cost artificial intelligence large model, the release of XVERSE-MoE-A4.2B will undoubtedly have a significant impact on various industries. In the future, we have every reason to believe that YuanXiang will continue to delve into the field of artificial intelligence and introduce more groundbreaking AI technologies, promoting the development of China’s AI industry.
【来源】https://mp.weixin.qq.com/s/U_ihKmhRD6Xc0cZ8hMJ1SQ
Views: 5