Alibaba Cloud announced a series of AI model and tool upgrades at its 2025 Spring Launch event, aimed at bolstering its international offerings and reaffirming its commitment to global AI innovation.
The latest offerings include expanded access to Alibaba’s large language model (LLM) series, Qwen, through the company’s Singapore availability zones. The newly available models include the large-scale Mixture of Experts (MoE) model Qwen-Max, the reasoning model QwQ-Plus, the visual reasoning model QVQ-Max and the end-to-end multimodal model Qwen2.5-Omni-7b.
QwQ-Plus is an advanced reasoning model designed for deep analytical thinking, tackling complex challenges such as expert-level math problems and intricate question-and-answer tasks with precise, algorithm-driven solutions. QVQ-Max excels in visual reasoning, effectively addressing multifaceted multimodal problems with high accuracy and supporting visual inputs while delivering chain-of-thought outputs.
Qwen, launched in April 2023, is one of the most popular AI model families worldwide. It has been adopted by over 290,000 customers across industries, from automotive, manufacturing, and finance to gaming, healthcare, and robotics. It is accessible via Alibaba Cloud’s generative AI development platform, Model Studio.
“We are launching a series of Platform-as-a-Service（PaaS）and AI capability updates to cater to the growing global demand for digital transformation. These enhancements will deliver more secure and high-performance services that enable businesses to scale and innovate in an AI-driven environment,” said Selina Yuan, President of International Business at Alibaba Cloud Intelligence.
PaaS Enhancements to Better Support AI Solutions
On the PaaS front, Alibaba Cloud’s Platform for AI (PAI) introduced new functionalities to deliver scalable, cost-effective, and user-friendly solutions for generative AI and LLMs. The new PAI-Elastic Algorithm Service (EAS) features distributed inference capabilities with a multi-node architecture, catering to the needs of super-large models with MoE structure and ultra-long-text processing. A prefill-decode disaggregation function was also introduced, resulting in a remarkable 92% increase in concurrency and a 91% boost in tokens per second (TPS) when paired with the Qwen2.5-72B model.
The upgraded PAI-Model Gallery now offers a comprehensive selection of nearly 300 advanced open-source models, including the entire range of Alibaba Cloud’s proprietary Qwen and Wan series. These models are accessible through a user-friendly, no-code deployment and management interface, which now includes new features such as model evaluation for performance insights and model distillation to reduce deployment costs.
To enhance efficiency in data management, Alibaba Cloud has enhanced its flagship cloud-native relational database, PolarDB, with Qwen-powered AI inference capabilities. This integration enables in-database machine learning capabilities, eliminating the need for data shifting in inference workflows and significantly reducing processing latency while enhancing efficiency and data security.
It also integrated its data warehouse AnalyticDB into Model Studio as the recommended vector database for RAG solutions. This enhancement connects enterprise proprietary knowledge bases directly to AI models and tools available on Model Studio, streamlining the development of context-aware applications.
SaaS AI Tools to Accelerate Industry Transformation
In the realm of SaaS, Alibaba Cloud launched an array of new Software-as-a-Service (SaaS) AI products. Highlights include AI Doc, an intelligent document processing tool that leverages LLMs for accurate parsing and information extraction from documents, and Smart Studio, an AI-powered content creation platform that enhances marketing and creative outputs through text-to-image, image-to-image and text-to-video applications.
Alibaba Cloud also unveiled SmartQ, an AI-driven data analysis module within its Quick Business Intelligence (BI) platform. This chat-based BI solution facilitates non-technical users with AI to generate insights and visualize key business data by posing questions in plain language.
Additionally, a new AI search function has been added to Alibaba Cloud’s official website to assist prospective customers, especially SMEs, discover solutions and insights for strategic decision-making.
In addition, Alibaba Cloud announced upgraded incentive policies for resellers and distributors, including higher commission rates and support resources, aiming to foster growth and collaboration within its ecosystem.
Earlier this year, Alibaba Group declared a historic investment of RMB 380 billion (US$ 53 billion) over three years to further develop its cloud computing and AI infrastructure. This investment surpasses Alibaba’s total AI and cloud expenditure over the past decade and reinforces the company’s dedication to AI-driven growth as a leading global cloud provider.
