Qwen models are developed by Alibaba Cloud. Open source Qwen large language models (LLMs) support a large number of parameters, such as 1.8 billion, 7 billion, 14 billion, and 72 billion. Based on the transformer architecture, these models are pre-trained by using an ultra-large amount of data. The data for pre-training Qwen models covers a wide range of data types, including text, professional books, and code. Open source Qwen models are available on ModelScope. DashScope also provides API services that can be directly used for the aligned versions of open source Qwen models, including qwen-1.8b-chat, qwen-7b-chat, qwen-14b-chat, and qwen-72b-chat.
For more information about how to use API operations to call open source Qwen models, see the following topics: