Alibaba is redesigning its Artificial Intelligence chatbot Qwen as it races to catch up with tools such as OpenAI’s ChatGPT. The updated app is replacing the older Tongyi version and was made available on both leading app stores last Friday.
Describing Qwen in its app-store descriptions as the “most powerful official AI assistant for its models” and the primary way for users to test-drive its latest Qwen model. The company also intends to equip shopping agent-like features to assist shoppers on other platforms like Taobao, Bloomberg reported. Alibaba did not immediately respond to a request for comment.
Over the last couple of years, Alibaba has been working to increase the use of its Qwen models at a time when people around the world are clamouring for AI tools. Alibaba, with newer Chinese players like DeepSeek and Moonshot AI on its heels, has emerged as one of the country’s larger AI developers. It has also encouraged an open-source community by publishing its models for others to use and modify.
Alibaba has been trying to convert these models into regular sources of revenue, and the push seems to be working. In the June quarter, sales of its AI products rose by triple digits for an eighth straight quarter.
Meantime, Alibaba reduced the cost of using its largest model, Qwen3-Max, by almost half, the South China Morning Post reported. The trillion-parameter system debuted in September, with some of the most expensive pricing on Alibaba Cloud. The firm has now adjusted its lowest API prices from US$0.861 to US$0.459 per million input tokens, and from US$3.441 to US$1.836 per million output tokens. In addition, users who run batch workloads on the system during off-peak hours receive a 50 per cent discount.
The model recently won a cryptocurrency investment competition, competing against the top models in both the US and China. Its price reduction coincides with tougher competition in China’s model market. Several companies, among them Moonshot AI, Zhipu AI and MiniMax, have introduced new systems in recent months and have touted their speed, power consumption and relatively low prices.
China has already experienced multiple waves of price cuts across the broader artificial intelligence space. There have been previous battles between big model developers and the new competition over tools like coding tools. This week, the cloud unit of ByteDance, Volcano Engine, unveiled a new coding agent priced at 9.90 yuan (US$1.30) for the first month.
Businesses have also experimented with various tactics to lure customers. On Tuesday last week, Moonshot AI — backed by Alibaba — introduced a deal allowing new users to experiment with its Kimi K2 Thinking model for as little as 0.99 yuan. Users were encouraged to haggle with the Kimi chatbot for their own discount, sharing “prompt injection” tips from other users and online sources that presented messages as though they came from someone who claimed to work for Moonshot. Hours after it was activated, the company said that the chatbot had begun “hallucinating,” and engineers were sent in to set it straight.
The similar sprint by Alibaba in AI has not gone unnoticed in the United States. “Silicon Valley didn’t want to say it, but the symptoms are all too clear: we’re experiencing a large-scale Qwen panic,” marketing expert Tulsi Soni wrote on social media Saturday.
In the meantime, Alibaba had to fight back against allegations reported by the Financial Times. The report said that a White House memo accused Alibaba of offering the People’s Liberation Army technical assistance and access to customer data, including IP addresses, Wi-Fi details, payment information, and AI services. Employees of Alibaba had also leaked information about “zero-day” security vulnerabilities, the memo said.
The FT said it could not verify the memo independently, but that the assertions underscored increasing concern in Washington about threats posed by Chinese cloud and AI suppliers.
Alibaba denied the accusations, which it said were “entirely false,” and criticised the leak for its motives. A spokesman called the report a “malicious PR operation” intended to undermine a recent trade deal between the US and China.
A spokesman for the Chinese Embassy in the US also condemned FT’s reporting, describing the accusations as “baseless” and a misrepresentation of the facts.
