CX Briefing: Alibaba Launches New AI Models

CX Briefing: Alibaba Launches New AI Models


Published: Apr. 29, 2025  6:51 p.m.  GMT+8,  Updated: Apr. 29, 2025  5:23 p.m.


您的浏览器不支持 audio 标签。


to this article

1x

A rundown of the news making headlines in and around China:

Alibaba’s new AI: Alibaba Group Holding Ltd. released a new open-source large language model (LLM) series on Tuesday that it said performs better and is cheaper to get up and running than a competing artificial intelligence (AI) model from DeepSeek. The e-commerce giant touted the Qwen3 LLMs as having one-third of the total parameters of the DeepSeek-R1. Alibaba said the difference in the number of Nvidia H20 chips needed to run the models lowers deployment costs by up to 75% compared with DeepSeek-R1. The release shows how Chinese tech companies remain bent on one-upping each other in the competition over foundational AI models. Alibaba also announced Tuesday that it will invest more than 380 billion yuan over the next three years on cloud and AI hardware infrastructure, an amount exceeding its total investment over the past decade.

loadingImg

You’ve accessed an article available only to subscribers

VIEW OPTIONS


Share now and your friends
will read it for free!

Share this article

Open WeChat and scan the QR code



Originally Appeared Here