Alibaba Group Holding, the Chinese tech giant, recently announced that more than 90,000 corporate clients have adopted its self-developed large language models (LLMs).
This is a huge deal, as companies across various industries, from consumer electronics to cars and online games, are embracing Alibaba’s AI offerings.
Qwen: Alibaba’s LLM Series is Making Waves
Alibaba’s LLM series, known as Tongyi Qianwen or simply Qwen, is the star of the show here. Zhou Jingren, the technology chief at Alibaba Cloud, states that our enterprise customers truly embrace Qwen, and we have observed numerous creative applications of the models across various industries.
But that’s not all—more than 2.2 million corporate users also have access to Qwen-powered AI services through DingTalk, Alibaba’s office collaboration platform. It’s clear that Alibaba is making big strides in the AI space.
The Race for Generative AI Dominance
Ever since OpenAI launched its game-changing ChatGPT in late 2022, tech companies in China have been racing to develop and promote similar services. Alibaba launched Qwen just a year ago, hot on the heels of rival Baidu’s Ernie Bot, which boasts 85,000 corporate clients as of early April this year.
Qwen’s Innovative Applications
So, what exactly are these corporate clients using Qwen for? Chinese smartphone giant Xiaomi has integrated Qwen’s question-and-answer functionality into its smart assistant, Xiao Ai, which powers its handsets and even its new SU7 electric vehicle. Qwen also supports cool new features in Xiaomi’s latest smartphones, like analyzing food images and generating recipes.
Other major Qwen clients include Perfect World Games, a Beijing-based video game developer. It’s exciting to see how these companies are leveraging Alibaba’s AI capabilities to innovate and enhance their products and services.
Introducing Tongyi Qianwen 2.5 and Open-Source Models
Alibaba unveiled Tongyi Qianwen 2.5, an updated version of its LLM, at an event in Beijing on Thursday. The company claims that this new version performs better in various Chinese capabilities compared to GPT-4, which is OpenAI’s most advanced model available to the public.
In addition to Tongyi Qianwen 2.5, Alibaba also launched several open-source models of various sizes. Among these, the Qwen1.5-110B stands out due to its training with a staggering 110 billion parameters. For context, the majority of mainstream Chinese open-source models have undergone training with 7 billion to 14 billion parameters. Generally speaking, more parameters mean more complex and capable AI models.
The Future of AI in China
With Alibaba making such significant strides in the AI space, it’s clear that the competition among Chinese tech giants is heating up. As more and more companies adopt and integrate these advanced AI models into their products and services, we can expect to see some truly innovative applications emerge.
The Information is Taken from CNBC and MSN