Alibaba unveils new Qwen3.5 model for ‘agentic AI era’
BEIJING, Feb 16 (Reuters) – Alibaba announced a new launch on Monday. artificial intelligence The Qwen 3.5 model is designed to run complex tasks independently, delivering major improvements in performance and cost, and the Chinese tech giant claims to outperform major rival models in the US in several benchmarks.
The release comes as Alibaba aims to attract more users to its Qwen chatbot app in China, a landscape currently dominated by rival tech giant ByteDance’s Doubao and DeepSeek, the first Chinese AI firm to break ground globally last year.
Alibaba said Qwen3.5 is 60% cheaper to use and eight times better than its predecessor at handling large workloads, adding that the model also comes with the ability to perform actions independently across mobile and desktop apps, or what the company calls “visual agency capabilities.”
“Designed for the age of agentic AI, Qwen3.5 is designed to help developers and organizations move faster and do more with the same computing, setting a new benchmark for capability per unit inference cost,” the company said in a statement.
ByteDance on Saturday launched Doubao 2.0, an upgrade to its chatbot app, which now has the largest user base in China, approaching 200 million. The announcement, like Alibaba’s, positioned the new model as relevant for the age of artificial intelligence agents.
The launch of Qwen3.5 could help build on Alibaba’s recent gains in the fierce competition of AI models in China. Earlier this month, the e-commerce giant’s coupon giveaway campaign that encouraged consumers to buy food and drinks directly through the Qwen chatbot led to a sevenfold increase in active users despite some hiccups.
Last year, the e-commerce giant became one of the first DeepSeek competitors to respond to the startup’s viral rise by launching the Qwen 2.5-Max, which it claimed was superior to one of DeepSeek’s popular models.
The company didn’t mention DeepSeek in its Qwen3.5 announcement, and the few benchmarks it has released only show that the new model outperforms the previous iteration and rival US models GPT-5.2, Claude Opus 4.5, and Gemini 3 Pro.
DeepSeek is expected to launch its new generation model in the coming days; This is raising expectations among investors and industry insiders given the global tech stock selloff the company triggered a year ago.
(Reporting by Eduardo Baptista; Editing by Sam Holmes)
