Alibaba rolls out revamped Qwen chatbot as model pricing drops

Alibaba is updating its synthetic intelligence chatbot Qwen as it pushes to meet up with instruments like OpenAI’s ChatGPT. The revised app replaces the older Tongyi model and have become accessible on each main app shops on Friday, final week.

In its app-store descriptions, Alibaba calls Qwen the “strongest official AI assistant for its fashions” and the principle manner for customers to attempt its latest Qwen model. The firm additionally plans so as to add agent-style options that may assist consumers on platforms such as Taobao, in response to Bloomberg. Alibaba didn’t reply to a request for remark.

Alibaba has spent the previous two years making an attempt to develop the usage of its Qwen fashions throughout the international rush towards AI instruments. Along with newer Chinese gamers like DeepSeeokay and Moonshot AI, Alibaba has grow to be one of many nation’s larger AI builders. It has additionally supported an open-source method by making its fashions accessible for others to make use of and adapt.

Alibaba has been making an attempt to show these fashions into regular income, and the push seems to be paying off. In the June quarter, gross sales from its AI merchandise grew at triple-digit charges for the eighth quarter in a row.

At the identical time, Alibaba lower the price of utilizing Qwen3-Max, its largest model, by nearly half, in response to the South China Morning Post. The trillion-parameter system launched in September with a few of the highest costs on Alibaba Cloud. The firm has now decreased its lowest API charges from US$0.861 to US$0.459 per million enter tokens and from US$3.441 to US$1.836 per million output tokens. Users who run batch duties throughout off-peak hours get one other 50 per cent low cost.

The model not too long ago positioned first in a cryptocurrency funding contest that in contrast high fashions from each the US and China. Its value drop comes amid sharper competitors in China’s model market. Several start-ups — together with Moonshot AI, Zhipu AI, and MiniMax — have launched new programs in latest months and have promoted their efficiency and low prices.

China has already seen a number of rounds of value cuts throughout the AI sector. There have been earlier battles between main model builders, adopted by new competitors in areas such as coding instruments. This week, Volcano Engine, the cloud unit of ByteDance, launched a brand new coding agent for 9.90 yuan (US$1.30) for the primary month.

Companies have additionally tried new methods to draw clients. Last week on Tuesday, Moonshot AI — which Alibaba backs — rolled out a proposal that lets new customers attempt its Kimi K2 Thinking model for as little as 0.99 yuan. People had been inspired to barter their very own low cost with the Kimi chatbot, which led some customers to share “immediate injection” suggestions on-line, together with messages that tricked the system into pondering they labored for Moonshot. Within hours, the corporate mentioned the chatbot had began “hallucinating,” and engineers had been known as to repair the problems.

Alibaba’s quick progress in AI has additionally been seen within the United States. “Silicon Valley doesn’t wish to admit it, however the signs are apparent: we’re witnessing a full-blown Qwen panic,” advertising and marketing specialist Tulsi Soni wrote on social media on Saturday.

At the identical time, Alibaba needed to defend itself towards claims reported by the Financial Times. The report mentioned a White House memo alleged Alibaba had supplied China’s People’s Liberation Army with sure technical assist, together with entry to some buyer knowledge such as IP addresses, Wi-Fi particulars, fee info, and AI providers. The memo additionally claimed that Alibaba employees had handed alongside details about “zero-day” safety flaws.

The FT mentioned it couldn’t affirm the memo by itself however mentioned the claims pointed to rising concern in Washington in regards to the dangers tied to Chinese cloud and AI suppliers.

Alibaba rejected the allegations, saying they had been “fully false” and questioning the motives behind the leak. A spokesperson described the report as a “malicious PR operation” meant to weaken a latest commerce deal between the US and China.

A spokesperson for the Chinese Embassy within the US criticised the FT’s reporting as properly, calling the accusations “groundless” and a distortion of info.

(Photo by zhang hui)

See additionally: Alibaba’s new Qwen model to supercharge AI transcription tools

Want to study extra about AI and large knowledge from business leaders? Check out AI & Big Data Expo happening in Amsterdam, California, and London. The complete occasion is a part of TechEx and is co-located with different main expertise occasions, click on here for extra info.

AI News is powered by TechForge Media. Explore different upcoming enterprise expertise occasions and webinars here.

The publish Alibaba rolls out revamped Qwen chatbot as model pricing drops appeared first on AI News.