SEPTEMBER 9, NEWS. TODAY, THE WAVE SUMMIT IN-DEPTH LEARNING DEVELOPER CONFERENCE 2025, BEIJING。BaiduThe Chief Technical Officer, Director of the National Centre for Advanced Learning Technology and Applied Engineering Research, Wang Seamount, was officially releasedWenxin Large Model X1.1 deep thinking model.
IT WAS DESCRIBED THAT THE X1-DEEP THINKING MODEL WAS BASED ON THE IN-DEEP THINKING MODEL BASED ON G4.5 TRAINING AND X1.1 WAS UPGRADED AGAIN. THE MODEL IS SIGNIFICANTLY ENHANCED IN TERMS OF DE FACTO, COMMAND COMPLIANCE, INTELLIGENCE, ETC。
1AI notes that users are already able to use the Magna Caralog X1.1 on the Web of Words and Words. In addition, Wang Sea Peaks indicates that the Manning Heart model X1.1 is officially on line with the 100-degree smart cloud-and-show platform, which is fully accessible to business clients and developers。
ACCORDING TO A LIVE PRESENTATION BY WANG SEAMOUNTS, X1.1 IS BASED ON THE USE OF AN ITERATIVE, MIXED AND INTENSIVE LEARNING TRAINING FRAMEWORK, WHICH ENHANCES THE OVERALL EFFECTIVENESS OF THE MODEL, ON THE ONE HAND, THROUGH A COMBINATION OF ENHANCED LEARNING, WHILE ENHANCING THE EFFECTIVENESS OF GENERIC AND INTELLIGENT TASKS, AND ON THE OTHER HAND, THROUGH THE PRODUCTION AND TRAINING OF AN ITERATIVE BASIS OF SELF-EVAPORATED DATA. 34.8%, COMMAND FOLLOWS 12.5%, SMART 9.6%。
IT WAS DESCRIBED THAT MANSHIN X1.1 OUTPERFORMED OVERALL PERFORMANCE IN MULTIPLE BASELINE ASSESSMENTS DeepSeek R1-0528, showing a lead in some tasks. At the same time, the effects were equal to those of the top international models GPT-5 and Gemini 2.5 Pro。
On-site, the Conference also released version 3.2 of the core framework of the propellers, which was fully upgraded in large model training, hardware suitability and ecological support, and synchronized with the upgrade of the large model development suite ERNIEKit and the efficient deployment package FastDeploy. The number of service enterprises is 760,000。