Wenxin Large Model User Scale Reaches 45 Million, Over 500 Plugins


Baidu released the new generation Wenxin large model ERNIE-5.0-0110, ranking eighth with 1460 points on the LMArena global text ranking list, making it the only Chinese domestic large model to enter the top ten. Its ability to handle mathematical problems is particularly outstanding, rising to the second globally,仅次于GPT-5.2-High.
Baidu released the Wenxin Large Model ERNIE-5.0-0110, ranking eighth in the LMArena Global Text Ability Ranking, making it the only domestic model to enter the top ten. Its performance in the field of mathematical reasoning is particularly outstanding, ranking second globally, just behind the unreleased GPT-5.2-High, demonstrating breakthroughs of domestic large models in specialized fields.
Baidu's ERNIE-4.5-21B-A3B-Thinking, a lightweight MoE model, tops Hugging Face's text generation chart. With 21B parameters (3B active/token), it offers efficient reasoning and 128K context support, marking a milestone in China's AI ecosystem.....
Today, Baidu App announced a major update, officially launching a completely new AI entry, and deeply integrating the Wenxin large model with the capabilities of DeepSeek to provide users with an upgraded experience. The core highlights of this upgrade include the integration of several cutting-edge intelligent features, such as 'AI Search', 'Web Information Refinement', 'Squeeze AI Summary', and 'Digital Human Intelligence'. These features enable the Baidu App to better understand user needs and offer more personalized search results.
Recently, Baidu Search announced the full integration of the latest deep search features from DeepSeek and the Wenxin Large Model, aimed at providing users with a more diversified and intelligent search service. According to official sources, the deep search feature of the Wenxin Large Model was officially launched on February 13. With its robust reasoning and planning capabilities, as well as tool invocation abilities, it can deliver expert-level content responses and effectively address task requirements in various scenarios, enabling multimodal input and output, further enriching the user's search experience.