Alibaba Cloud Launches AI Agent Tools and Model Updates for Developers
Alibaba Cloud announced a suite of products and infrastructure upgrades for building AI agents at its first international Qwen Conference in Singapore on May 26. The company released Qwen3.7-Max, a new large language model, alongside cloud infrastructure designed to let AI agents call cloud services as easily as functions.
The announcements target product teams and developers building applications with AI agents. Alibaba Cloud positioned these tools as the infrastructure layer needed for what it calls the "agentic era"-where AI systems autonomously execute tasks across cloud resources and applications.
Qwen3.7-Max ranks fifth globally
Qwen3.7-Max scored 56.6 points on Artificial Analysis's global large language model Intelligence Index, placing it fifth worldwide and first among Chinese models. It outperformed Kimi-K2.6, DeepSeek-v4-Pro-Max, and GLM5.1, with performance comparable to GPT, Claude, and Gemini.
The model is available now on Model Studio, Alibaba's AI development platform, in the Singapore region.
Cloud skills converted to agent-callable formats
Alibaba Cloud launched a Skills portal that converts capabilities from over 60 cloud products into formats that AI agents can invoke. The portal uses Skill-based and MCP-compatible formats, allowing agents to call database, data processing, operations, and security tools as function calls.
The company also upgraded its infrastructure to support agent runtime environments with lightweight sandboxes, cross-task memory, data circulation, and automated operations management across its stack.
Qwen Cloud platform simplifies model deployment
Qwen Cloud is a new platform designed to make it easier for developers and enterprises to build AI applications and agents. It offers three entry points: a Skills interface for agents, a command-line interface for workflow integration, and a website for human users.
The platform aggregates Alibaba's Qwen models, open-source models, and third-party offerings for text, vision, audio, image, video, and embedding tasks. This approach lets product teams choose models without managing multiple vendor integrations.
Enterprise agent toolkits for teams
Alibaba Cloud released the JVS Agent Suite, a set of tools for building and running AI agents at scale. JVS Claw Teams supports 24/7 cloud operations, centralized distribution of proprietary skills, and integrated security management.
JVS Mobile, another suite component, enables AI agents to work across mobile applications with multi-agent collaboration and complex task execution. Both products are built on the OpenClaw framework with cloud-native security.
Singapore initiative targets 1,000 workers and students
Alibaba Cloud partnered with Tech Talent Assembly and ST Telemedia Global Data Centres to train over 1,000 SMEs and students in Singapore on generative and agentic AI. The program provides access to Alibaba's tools alongside hands-on training.
Desmond Tan Kok Ming, Senior Minister of State in Singapore's Prime Minister's Office, said the initiative reflects efforts to support workers in the AI economy through the government's AI-Ready SG initiative.
Developer competitions and open-source commitment
Alibaba Cloud launched a global hackathon for developers to build production-grade AI agents using Qwen models on Qwen Cloud. A separate film competition with Picsart invites creators to produce AI-generated films using HappyHorse, Alibaba's video generation model.
The company also joined the PyTorch Foundation as a Platinum member, committing resources to open-source AI infrastructure development.
For product development teams, the significance lies in the infrastructure layer: these tools reduce the friction of integrating cloud services into agent workflows. Rather than writing custom API integrations, developers can let agents call cloud capabilities directly through standardized formats.
Your membership also unlocks:
