Qianwen QwQ-32B is the latest powerful reasoning model released by Alibaba, which is characterized by both performance and efficiency. The Qianwen QwQ-32B has achieved a qualitative leap in mathematics, coding and general capabilities, and its overall performance is comparable to that of DeepSeek-R1.
Tongyi App launches Qianwen QwQ-32B agent, and the AI experience continues to be upgraded
On March 6, the reporter discovered that Tongyi, an AI application owned by Alibaba’s AI To C business, had been connected to Alibaba’s latest open source reasoning model Tongyi Qianwen QwQ-32B, and is currently available on mobile and web terminals.
Qianwen QwQ-32B is the latest powerful reasoning model released by Alibaba, which is characterized by both performance and efficiency. The Qianwen QwQ-32B has achieved a qualitative leap in mathematics, coding and general capabilities, and its overall performance is comparable to that of DeepSeek-R1. While maintaining strong performance, Qianwen QwQ-32B also significantly reduces deployment costs and can also be deployed locally on consumer-grade graphics cards.
Users can experience it through the Qwen-QwQ-32B agent portal on Tongyi App and Tongyi web version.
Previously, Tongyi App has also launched the Qwen-QwQ-preview version of the reasoning model in the form of an agent, allowing domestic users to experience Tongyi Qianwen’s powerful model capabilities as soon as possible. Some netizens pointed out on social media that its code generation ability performs well and generates macros basically once.
It is understood that in the future, Tongyi App and Tongyi web version will continue to launch Tongyi Thousand Questions ‘latest model capabilities, and apply in-depth reasoning capabilities to its agent ecosystem to innovate the personal AI assistant experience and further enhance users’ lives, work, and learning efficiency.
In addition, the reporter also found that the product interaction of Tongyi App is also quietly changing. The home page has become a cute girl next door as the main interaction object, and many practical functions such as taking photos and teaching questions have been added. The original agent ecosystem has also been further enriched.