Alibaba unveiled its latest artificial intelligence model, Qwen 3.5, on Monday, positioning the upgrade as purpose-built for an emerging era of autonomous AI agents capable of executing complex tasks independently.
The release comes as Chinese technology giants jostle for dominance during the Lunar New Year holiday period, one year after startup DeepSeek upended the global AI landscape with its low-cost models. The Hangzhou-based e-commerce giant said Qwen 3.5 is 60 percent cheaper to operate and eight times better at processing large workloads than its predecessor.
The model features what Alibaba describes as “visual agentic capabilities,” enabling it to take autonomous actions across mobile and desktop applications without user intervention. Alibaba stated that Qwen 3.5 is “designed to help developers and enterprises move faster and do more with the same compute, setting a new benchmark for capability per unit of inference cost.”
The company claimed the new model outperforms its earlier version and beats U.S. rivals including OpenAI’s GPT-5.2, Anthropic’s Claude Opus 4.5, and Google’s Gemini 3 across several performance benchmarks. The model also supports more than 200 languages, including many used in South Asia, Oceania, and Africa, alongside enhanced reasoning and improved image and video processing capabilities.
The release intensifies competition in China’s AI market, where Alibaba trails ByteDance’s Doubao chatbot. According to QuestMobile data published in late December, Doubao leads Chinese AI chatbot apps with 155 million weekly active users, followed by DeepSeek with 81.6 million.
ByteDance launched Doubao 2.0 on Saturday, also framing its upgrade as suited for the “agent era.” Alibaba has recently gained ground through aggressive marketing. Earlier this month, the company’s coupon giveaway campaign—part of a 3-billion-yuan ($433 million) promotional push—enabled users to purchase food and beverages directly through the Qwen chatbot, leading to a seven-fold increase in active users despite some technical glitches that temporarily overwhelmed the system.
The flurry of model releases comes ahead of an anticipated new model from DeepSeek, whose cost-efficient approach a year ago challenged assumptions that only companies spending tens of billions on computing infrastructure could produce cutting-edge AI. The current frenzy traces back to DeepSeek’s breakthrough during the 2025 Spring Festival, when its R1 reasoning model triggered a surge in user adoption and briefly overtook Doubao in downloads.
That success demonstrated how a single viral moment could reshape market dynamics, prompting established players to pour resources into both technology and marketing.
