JAKARTA - On Thursday 3 August, Alibaba Group, a Chinese technology and e-commerce company, announced the launch of two artificial intelligence (AI) models with open sources from their cloud computing department.

Two large language models (LLM) named Qwen-7B and Qwen-7B-Chat each have 7 billion parameters. Alibaba stated that these two models are small size versions of Tongyi Qiawen, which the company released in April.

These new models aim to help introduce AI into small and medium business operations.

The company states that Qwen-7B and Qwen-7B-Chat have a variety of capabilities of interest to companies, such as "codes, model weights, and documentation that will be freely accessible to academics, researchers, and commercial institutions around the world."

Alibaba's latest NGO is also the first to be released from a technology company from China with open sources. However, the company states that businesses with more than 100 million monthly active users will require a license.

On August 1, the company also announced an update in the form of a vector machine for its warehousing AnalyticDB data service, which allows its corporate clients to quickly create custom-generative AI applications.

The development comes after Meta released their LLM with open sources Llama 2 together with Microsoft on July 16.

Meta says that Llama 2 they are trained by using 40% more public data and can process twice the context over its predecessor. The model is also an open source, with the largest version of Llama 2 having 70 billion parameters.

Like the latest model from Alibaba, this model requires a license from the company with more than 700 million monthly active users.

On July 26, Alibaba announced "the first solution to training and implementation for the entire Llama2 series in China" after implementing the Llama 2 solution for businesses in the development of AI-powered software and tools.


The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)