TLDR: Alibaba Cloud has significantly upgraded its Bailian platform at the 2025 Yunqi Conference, introducing the ModelStudio-ADK high-code framework. This initiative aims to address the limitations of low-code AI Agent development by enabling enterprises to build robust, autonomous, and commercially viable AI Agents for complex business scenarios. The platform offers a comprehensive stack from models to infrastructure, supporting over 200,000 developers who have created over 800,000 Agents in the past year, demonstrating a 15-fold increase in monthly model invocation volume.
At the 2025 Yunqi Conference, held on September 24th, Alibaba Cloud unveiled a significant upgrade to its Bailian platform, introducing the ModelStudio-ADK high-code framework. This strategic move aims to revolutionize enterprise AI Agent development by overcoming the inherent limitations of traditional low-code solutions, which often struggle with complex tasks, cross-system integrations, and multi-round autonomous decision-making.
Zhou Jingren, Chief Technology Officer of Alibaba Cloud Intelligence Group, underscored the importance of this shift, mentioning “Agent” 38 times and “intelligent agent” 33 times during his address. He articulated a vision where AI Agents move beyond simple Q&A and process automation to become integral, stable, and scalable components of business operations. Wu Yongming, CEO of Alibaba Group, further emphasized this long-term perspective, stating, “Artificial General Intelligence (AGI) is not the end of AI development but a brand-new starting point. AI will not stop at AGI; it will move towards Artificial Superintelligence (ASI), which surpasses humans and can self-iterate and evolve.” He positioned Agents as a crucial step towards achieving ASI and unlocking practical value in various industries.
ModelStudio-ADK is designed to empower enterprises to develop Agents with advanced capabilities such as autonomous planning and cyclic execution. It integrates a comprehensive suite of seven enterprise-level functionalities, including sophisticated tool invocation, intelligent memory access, dynamic reasoning scheduling, a secure sandbox environment, full-link observability, and a pioneering commercial payment channel developed in collaboration with Alipay. This dual-track approach, combining low-code ADP for rapid prototyping and high-code ADK for in-depth customization and large-scale deployment, provides Alibaba Cloud with a competitive edge, focusing on “high-code and strong capabilities” for production-ready Agents.
Alibaba’s commitment to a full-stack AI ecosystem underpins Bailian’s capabilities. This includes the Tongyi series of large models at the top, the Bailian enterprise-level development and operation platform in the middle, and robust cloud computing infrastructure (computing power, storage, databases) at the bottom. This integrated chain ensures that Agents are not only quickly developed and stably operated but also continuously iterated with technological advancements.
The platform has already demonstrated significant traction. In the past year, over 200,000 developers on Alibaba Cloud’s Bailian platform have created more than 800,000 Agents, leading to a 15-fold increase in monthly model invocation volume. A PwC survey in May, involving 300 global executives, revealed that 79% of companies are already using AI Agents, reporting benefits such as 66% increased productivity, 57% cost reduction, 55% faster decision-making, and 54% improved customer experience.
ModelStudio-ADK’s impact is evident in real-world applications. For instance, MYbank developed a loan review application using ADK that can process 26 types of vouchers and identify over 400 fine-grained objects with an accuracy exceeding 95%. This has dramatically reduced a 3-hour manual process to less than 5 minutes, showcasing the framework’s value in complex financial scenarios.
Xu Zhiyuan, a senior product expert at Alibaba Cloud’s Bailian, summarized ADK’s essence: “It defines various core capabilities and components of Agents to meet the construction logic under complex task scheduling through a code-based approach combined with the capabilities of models.” The framework supports over 200 mainstream models, offering flexibility between performance and controllability, and provides standardized services for tool invocation, knowledge base retrieval, memory management, and sandbox environments.
Also Read:
- Alibaba’s Qwen Team Unveils Qwen3-Coder: A 480 Billion Parameter Open-Source AI for Advanced Coding
- Moonshot AI’s Kimi Chatbot Unveils ‘Agent Mode’ for Automated Website and Presentation Creation
In the competitive landscape of AI Agent platforms, where major players like LangChain/LangGraph, Microsoft AutoGen, Google ADK, Tencent, and Huawei are also making strides, Alibaba Cloud’s integrated “platform + framework + components” strategy aims to provide a comprehensive solution from development to deployment, operation, and commercialization. As OpenAI CEO Sam Altman predicted earlier this year, “in 2025, we may see the first AI agents ‘join the workforce’ and materially change the output of companies.” Alibaba Cloud’s Bailian platform is actively working to make this a reality.
