Alibaba Cloud has announced several enhancements to its AI offerings during its annual developer summit.

Among the updates are new large language models (LLMs), advanced AI development tools, upgraded cloud infrastructure, and a dedicated program to support global developers.

“Alibaba Cloud is committed to delivering real value to global developers through cutting-edge AI models, enhanced cloud infrastructure, and accessible support programs,” said Dongliang Guo, VP of International Business and Head of International Products and Solutions at Alibaba Cloud Intelligence. 

“Together, we aim to spark more AI-driven innovations, benefiting startups, enterprises, and industries altogether across the globe.”

Alibaba expands Qwen AI model suite

Central to Alibaba’s efforts is the expansion of its proprietary Qwen LLM family. The Qwen 2.5 series – available in parameter sizes from 7 billion to 72 billion – is now accessible via APIs on the company’s generative AI development platform, Model Studio. These models support applications such as text generation, language understanding, and translation.

But Alibaba Cloud hasn’t stopped with text-based models. It has added capabilities through multimodal AI models:

• Qwen-VL series: Integrates advanced vision understanding for tasks like image analysis.  
• Wanx2.1 (known locally as Tongyi Wanxiang): A visual generation model for creating images from text prompts.  
• Qwen-Audio: Specialises in audio language applications.  

Developers can now also access Tongyi Lingma, an AI coding assistant powered by Qwen 2.5-coder. The tool provides features such as code completion, debugging assistance, batch unit test creation, and code snippet searches to streamline the software development process and improve productivity.

Additionally, Model Studio now offers new AI development tools designed to enhance reliability and scalability for global developers. Highlights include:

• Workflow: Breaks down complex tasks into subtasks to improve organisational control.  
• Agent: Enables multi-agent collaboration for planning and execution.  
• RAG (Retrieval-Augmented Generation): Augments model accuracy using external data sources.  
• Batch Reasoning: Processes multiple prompts simultaneously for better efficiency.  
• AutoEval (Automated Model Evaluation): Assists in assessing models for optimal deployment.  

These tools, along with deployment support and observability services, will be fully available by the end of this month.

Infrastructure upgrades  

Powering these AI advancements requires robust infrastructure. To that end, Alibaba Cloud has introduced its 9th Generation Enterprise Elastic Compute Service (ECS) instances that are set to debut in global markets this April.

Compared to its predecessors, the new ECS instances offer a 20% boost in computing efficiency. Features like eRDMA (elastic Remote Direct Memory Access) improve support for intensive workloads such as high-performance computing and search recommendations by up to 50%.

Another major addition, Alibaba Cloud Container Compute Service (ACS), is set to launch internationally in January 2025.

By integrating container technology with underlying cloud resources, ACS reduces operational complexity and costs. This innovation is designed to let developers focus on building applications rather than wrestling with infrastructure management.  

GenAI Empowerment Program

To cultivate a thriving generative AI community, Alibaba Cloud has rolled out the GenAI Empowerment Program. Designed specifically for startups and developers utilising its Qwen models, the initiative offers several benefits:

• Free cloud credits to support development.  
• Invitations to training workshops, tech expos, and demo events to refine skills and showcase innovations.  
• Opportunities for co-marketing to enhance visibility within global ecosystems.  

The program is intended as a launchpad for generative AI projects, connecting participants with a network of like-minded innovators while reducing entry barriers to advanced AI tools.

Several international organisations are already leveraging Qwen-based solutions to drive progress in distinct sectors.

Japanese AI company Axcxept has adapted the Qwen 2.5 LLM to create EZO, a lightweight AI model designed to handle tasks such as coding, reasoning, roleplay, and complex writing in Japanese. The model provides low-latency, reliable performance and has become particularly useful in healthcare and public sector applications.

“Qwen 2.5 has significantly enhanced its ability to process Japanese, giving it a competitive edge over other models,” said Kazuya Hodatsu, CEO of Axcxept. “With Axcxept’s proprietary training process, we have developed a Japanese LLM that delivers unmatched accuracy.”

OxValue.AI, a deep-tech spinoff from the University of Oxford, employs Qwen-based multimodal AI models for corporate valuation. By analysing diverse data sources – such as text and audio related to a company’s R&D and financial operations – OxValue provides precise, cost-efficient assessments tailored to its customers.

“Processing diverse data sources is essential for our valuation services,” explained Professor Xiaolan Fu, Founder of OxValue.AI. “With the support of Alibaba Cloud’s AI models, we’ve significantly improved the quality and efficiency of this process. By collaborating with Alibaba Cloud, we’re able to deliver greater value to our corporate clients.”

As the AI arms race intensifies, Alibaba Cloud’s holistic approach may help it stand out—offering value to developers, startups, enterprises, and researchers alike. By reducing technical and financial obstacles, Alibaba aims to ignite innovation across industries.

(Photo by Zhang Hui)

See also: Mistral AI sets code generation benchmark with Codestral 25.01

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is co-located with other leading events including Intelligent Automation Conference, BlockX, Digital Transformation Week, and Cyber Security & Cloud Expo.

Explore other upcoming enterprise technology events and webinars powered by TechForge here.