%20--%3e%3cdefs%3e%3cstyle%3e%20.st0%20{%20fill:%20%23061b40;%20}%20.st1%20{%20fill:%20%23306af1;%20}%20.st2%20{%20fill:%20%235ce5cf;%20}%20%3c/style%3e%3c/defs%3e%3cg%3e%3cpath%20class='st0'%20d='M55,10.5h9v3h-9v9h12V7.5h-12v3ZM64,19.5h-6v-3h6v3Z'/%3e%3cpolygon%20class='st0'%20points='69%2016.5%2078%2016.5%2078%2019.5%2069%2019.5%2069%2022.5%2081%2022.5%2081%2013.5%2072%2013.5%2072%2010.5%2081%2010.5%2081%207.5%2069%207.5%2069%2016.5'/%3e%3cpolygon%20class='st0'%20points='95%2010.5%2095%207.5%2083%207.5%2083%2022.5%2095%2022.5%2095%2019.5%2086%2019.5%2086%2016.5%2095%2016.5%2095%2013.5%2086%2013.5%2086%2010.5%2095%2010.5'/%3e%3cpath%20class='st0'%20d='M40,1.5v21h11.6l1.4-1.4v-7.6h0c0,0-1.4-1.5-1.4-1.5l1.4-1.4V2.9l-1.4-1.4h-11.6ZM50,19.5h-7v-6h7v6ZM50,10.5h-7v-6h7v6Z'/%3e%3c/g%3e%3cpath%20class='st1'%20d='M23.1,24L14.7,4.8l-4.9,11.2h3.8l-1.8,4H3.3L12.1,0H2C.9,0,0,.9,0,2v20c0,1.1.9,2,2,2h21.1Z'/%3e%3cpath%20class='st2'%20d='M34,0h-16.8l10.6,24h6.2c1.1,0,2-.9,2-2V2C36,.9,35.1,0,34,0ZM32.5,20h-4V4h4v16Z'/%3e%3c/svg%3e)
Tongyi Lab of Alibaba Cloud announced today that Tongyi DeepResearch has been fully open-sourced. This groundbreaking web proxy tool, with a lightweight scale of only 30 billion parameters (actually activated 3 billion), has achieved a breakthrough in performance, matching the deep research tools of OpenAI.
Tongyi DeepResearch has shown excellent performance in multiple authoritative benchmark tests, including 32.9 points in Humanity's Last Exam, 45.3 points in BrowseComp, and 75.0 points in xbench-DeepSearch, marking a significant progress in complex information retrieval and reasoning fields for open-source AI agents.
This proxy is designed to handle multi-step, long-term deep information gathering tasks, supporting an ultra-long context of 128K. It can efficiently decompose complex questions, perform multi-step reasoning, and integrate massive data. Its technological innovations include continuous pre-training of agent data and on-policy reinforcement learning, which makes it show strong stability and reliability in scenarios such as multi-day itinerary planning and legal document analysis.
Tongyi DeepResearch offers two reasoning modes: the single model ReAct mode suitable for lightweight tasks, and the Heavy mode that supports parallel collaboration of multiple Agents and computing power expansion. Currently, this technology has been applied in practice, for example, integrated into Amap to help users plan personalized multi-day itineraries; at the same time, it empowers the legal assistant "Tongyi Farsi," significantly improving the efficiency of contract risk assessment and case analysis.
As an important part of the Tongyi series large model ecosystem of Alibaba Cloud, the open-sourcing of Tongyi DeepResearch will greatly reduce the development threshold of AI agents. The series has previously open-sourced more than 200 models of different parameter scales, and has generated over 100,000 secondary developed models on the Hugging Face platform. Experts believe that this open-sourcing will provide developers and small teams with a powerful foundation, helping them quickly build AI agents in specific fields.