On November 17, 2025, Alibaba officially announced the launch of its “Qwen” project. The public beta version of the Qwen App was simultaneously released on major app stores, with its web and PC versions also made available to the public. The international version is set to debut in the near future. This marks another crucial move in Alibaba’s AI strategic layout following its announcement earlier this year of a 380 billion yuan investment in AI infrastructure. Regarded by the core management as “a decisive battle for the future in the AI era”, this launch signifies Alibaba’s full-scale entry into the AI-to-C market, initiating direct competition with ChatGPT.
The core competitiveness of the Qwen App stems from its robust technological underpinnings. Built on Qwen3, the world’s top-performing open-source model, its flagship variant Qwen3-Max has secured a spot among the top three models globally, outperforming international counterparts such as GPT-5 and Claude Opus 4. Since going fully open-source in 2023, the Qwen series has experienced rapid development. It has not only surpassed Meta’s Llama to become the world’s leading open-source model family but also gained extensive recognition across industries. Up to now, the global downloads of the Qwen series have exceeded 600 million, encompassing over 300 open-source models that cover multimodal capabilities including text, programming, and images, as well as a full scale ranging from 0.5B to 480B parameters. The number of derivative models has even exceeded 170,000. On the rankings of Hugging Face, a global AI open-source community, 7 out of the top 10 models worldwide are from Qwen. Its influence has extended to Silicon Valley—Airbnb relies heavily on Qwen models, and Jensen Huang, CEO of NVIDIA, has confirmed that Qwen holds the majority of the global open-source model market share, which continues to grow steadily.
The latest AI news shows that xAI has recently launched Grok 4.1, which has achieved remarkable results by topping the LMArena Text Arena rankings and significantly reducing factual hallucination rates, while Keda Xunfei has rolled out Spark X1.5 with performance reaching over 95% of GPT-5 . Against this backdrop of intensifying global AI competition, “being conversational and practical” serves as the core positioning of the Qwen App and a key factor in its differentiated competition. Even in its initial version, the app has demonstrated strong practical capabilities. With a single instruction, it can complete a research report and generate an elaborate PPT of dozens of pages within seconds. It even claimed the championship in a real – trading investment contest competing against world – class models like ChatGPT and Gemini. Alibaba plans to integrate diverse daily life scenarios such as mapping, food delivery, ticket booking, office work, shopping, and healthcare into the app in the future. It aims to build a service closed – loop featuring “recognition – recommendation – redirection”, creating a primary gateway for future AI – powered daily life and transforming AI from an entertaining interactive tool into a productivity tool integrated into daily life and production.
The launch of the Qwen App is no coincidence but an inevitable outcome of Alibaba’s 16 years of in-depth technological accumulation. In 2009, the establishment of Alibaba Cloud laid a world – leading cloud computing infrastructure for AI development. In 2017, the Dharma Institute was founded to focus on AI research and development. The company kicked off large – model R & D in 2018, rolled out the trillion – parameter large model M6 in 2021, and officially released its self – developed large model in September 2022. In September 2023, Alibaba established the core strategy of “user – first, AI – driven”. At the Cloud Summit held in September 2025, CEO Wu Yongming put forward the long-term goal of advancing towards “Artificial Superintelligence (ASI)” and upgraded the full-stack AI infrastructure. Continuous investment in underlying technologies has yielded remarkable results. Alibaba has become the only full-stack AI enterprise in China with the capability of software – hardware integration and vertical integration. The revenue of its AI – related products has maintained triple-digit year – on – year growth for eight consecutive fiscal quarters, driving Alibaba Cloud’s revenue to achieve its highest growth rate in three years. Meanwhile, the company’s stock price has risen by more than 87% since the beginning of the year.
In the view of Alibaba’s management, the current large – model sector has entered the Agentic AI era focused on “assisting humans”, making it an opportune time to enter the consumer market. The public beta of the Qwen App represents a significant implementation of Alibaba’s AI strategy. By adopting the two core approaches of “building the Android of the AI era through open – source and openness” and “constructing a super AI cloud”, Alibaba is leveraging its full-stack advantages to make strides in the AI sector. This not only marks the first time that a Chinese open-source model has launched direct competition against global leading AI products at the application level but also ushers in a new narrative of an open, professional, and practical full-stack AI ecosystem, which will accelerate the transformation of the AI industry from technological demonstration to practical implementation.
