Alibaba Cloud has unveiled a comprehensive range of models under the Qwen series housed within the Alibaba Cloud Model Studio. These models cater to a variety of uses, including text generation, embedding, and multi-modal applications. The latest in the series, Qwen2.5, offers enhanced capabilities building on its predecessors, highlighting improvements in processing instructions, generating long texts, and handling structured data such as tables.
The newly introduced Qwen2.5 models, available in both commercial and open-source variants, cater to specialized fields with sizes ranging from 7 billion to 72 billion parameters. This series encompasses Qwen-Max, Qwen-Plus, and Qwen-Turbo models, each optimized for distinct balances between performance, speed, and cost, along with specialized models such as the Qwen-VL for visual processing.
Moreover, Qwen models support a broad language base, including over 29 languages like Chinese, English, French, and Spanish. The models integrate enhanced expertise for specific fields, boosting coding and mathematical capabilities while supporting diverse system prompts. They also introduce modes for visual understanding and high-resolution image processing, suited for tasks such as object recognition, text analysis, and video comprehension in financial and business scenarios.
