Alibaba Cloud Unveils New AI Models and Revamped Infrastructure for AI Computing

Alibaba Cloud, the technology backbone of Alibaba Group, has released over 100 new large language models, Qwen 2.5, to the global open-source community. This was announced at the Apsara Conference. The Qwen models range from 0.5 to 72 billion parameters and support over 29 languages, catering to diverse AI applications across multiple sectors.

Alibaba Cloud has also introduced a revamped infrastructure aimed at AI computing needs. This new setup includes innovative cloud products enhancing computing, networking, and data center architecture. The goal is to support the development and application of AI models.

Alongside, Alibaba unveiled a new text-to-video AI model within the Tongyi Wanxiang family, capable of generating high-quality videos from text inputs. They also announced the Qwen2-VL model for better video comprehension and an AI Developer assistant for programming tasks.

Key upgrades to their AI infrastructure include the next-gen data center, CUBE DC 5.0, and the Alibaba Cloud Open Lake to optimize data utility. The PAI AI Scheduler improves computing resource management, and DMS for unified metadata management boosts data efficiency.

The 9th Generation Enterprise Elastic Compute Service (ECS) instance was also introduced, promising performance enhancements. These updates aim to provide comprehensive support, maximizing technology benefits for efficient, sustainable AI applications.