On February 21, Baidu's Wenxin Intelligent Agent platform announced the full integration of the DeepSeek model, which is now available to all users for free. The Wenxin Intelligent Agent platform is designed to help users quickly create and deploy intelligent agent applications based on the Wenxin large model.
The integrated DeepSeek model consists of four versions, covering various complex tasks and scenarios to meet different user needs. Among them, DeepSeek-R1 specializes in efficiently solving complex tasks, boasting excellent reasoning capabilities and fast inference speed, making it suitable for high-precision requirements. DeepSeek-R1-32B and DeepSeek-R1-14B are distilled versions of DeepSeek-R1, aimed at professional and high-performance demands, particularly excelling in complex tasks such as mathematical reasoning, code generation, and multi-step deductions. DeepSeek-V3 is a new generation general-purpose base model, pre-trained on a trillion-level dataset, supporting accurate associations with long contexts and multi-step logical reasoning, suitable for open-domain complex problem solving and interdisciplinary knowledge integration.
Currently, users can freely choose the desired DeepSeek model when creating intelligent agents on the Wenxin Intelligent Agent platform, and all users can experience these models at no cost. Although these models do not yet support online searches, knowledge base access, or plugin calls, their powerful reasoning abilities and multi-domain knowledge integration have already provided users with a rich array of application scenarios, catering to needs from complex reasoning and code generation to creative writing and emotional companionship—all in one place.
To help users get started quickly, the Wenxin Intelligent Agent platform offers a very simple operation process, allowing users to access the DeepSeek model with just one click.