EAI-100 Reveals Ant Lingbo Among Top Breakthroughs and Pioneers

At the "Zhuijin Jinling · AI Open Source Talent Summit and Moba Community Developer Conference" on March 22nd, the EAI-100 (Embodied Artificial Intelligence 100) list was officially unveiled. This annual recognition highlights the top 100 representative achievements and figures in embodied intelligence for 2025. The list was jointly released by the Moba Community alongside authorities including the CCF Intelligent Robotics Committee and the MIIT's Key Laboratory of Equipment Digital Twin Technology. Ant Lingbo Technology earned a dual distinction, being named in both the "Top 10 Breakthroughs of the Year" and the "Pioneer Figures 20" core lists.

The EAI-100 serves as a systematic annual review for the embodied intelligence field. Selection is based on criteria of tangible impact, enduring value, and directional contribution, assessing whether individuals and their work have driven substantial progress in research paradigms, system capabilities, or industrial implementation. This year's honorees include prominent figures and breakthroughs from institutions like Tsinghua University, Peking University, and The University of Hong Kong, as well as companies such as Yuzhu Technology, Galaxy General, Xinghai Map, and Zhiyuan Robotics.

(Pictured: Ant Lingbo Technology's Shen Yujun, selected for the EAI-100 Pioneer Figures list, alongside Yuzhu Technology's Wang Xingxing and Peking University's Wang He)

LingBot-VLA: A Cross-Task "General Brain" for Embodied Intelligence

Ant Lingbo Technology's self-developed embodied foundation model, LingBot-VLA, was selected as one of the "Top 10 Breakthroughs of the Year." The model delivers cross-modal and cross-task generalization for real-world robot operations, dramatically lowering post-training expenses and advancing the practical deployment of a "one-brain-for-many-machines" architecture.

LingBot-VLA is pre-trained on over 20,000 hours of real operational data spanning nine mainstream dual-arm robot configurations. It works in concert with Ant Lingbo's proprietary high-precision spatial perception model, LingBot-Depth, to further boost operational accuracy. The model requires only 80 demonstration samples to achieve high-quality task adaptation. Combined with deep optimizations to its underlying code library, it achieves a training throughput 1.5 to 2.8 times greater than mainstream frameworks, reducing both data and computational costs. Ant Lingbo has open-sourced LingBot-VLA and its associated post-training toolchain, empowering developers to efficiently adapt the technology to their specific scenarios and significantly enhancing its practical utility.

Chief Scientist Shen Yujun Named to the "Pioneer Figures 20"

Ant Lingbo Technology's Chief Scientist, Shen Yujun, was also honored in the "Embodied Intelligence - Pioneer Figures 20" list. This list recognizes individuals who have exerted a sustained and profound influence on the field, particularly acknowledging their role as "pathfinders" during critical phases of the discipline's evolution.

Dr. Shen Yujun, a graduate of The Chinese University of Hong Kong, has long focused on research in computer vision and generative models. He has published more than 100 papers in top-tier international conferences and journals such as CVPR and TPAMI, amassing over 10,000 citations. As the chief technical officer at Ant Lingbo Technology, he led the team in releasing and open-sourcing a comprehensive technical suite this January. This includes the spatial perception model LingBot-Depth, the foundation model LingBot-VLA, the world model LingBot-World, and the video-action model LingBot-VA. This portfolio builds a complete matrix from spatial perception to intelligent decision-making and from foundational capabilities to world modeling, accelerating the transition of embodied intelligence from laboratory research to large-scale, real-world application.