Model Releases

Ant Lingbo open-sources LingBot-Video, the world's first MoE video foundation model for embodied intelligence

Original title:全球首个面向具身智能的视频基模,蚂蚁灵波 LingBot-Video 开源

AI summary

Ant Lingbo Technology officially open-sourced LingBot-Video, the world's first MoE-architecture video generation foundation model for embodied intelligence. It has 30B total parameters but activates only about 3B during inference, improving efficiency by about 3x over Dense architectures of the same scale. The model incorporates 70,000 hours of robot data including VLA, VLN, and Ego, and aligns physical plausibility and task completion through a multidimensional reinforcement learning reward system. It scores 0.620 overall on RBench, surpassing models such as Wan2.6, and ranks first on the Physics-IQ Verified evaluation. It can be used for robot motion prediction, simulation data generation, and other directions.

Key Takeaways

  • Ant Lingbo Technology officially open-sourced LingBot-Video,the world's first MoE-architecture…
  • It has 30B total parameters but activates only about 3B during inference,improving efficiency by…
  • The model incorporates 70,000 hours of robot data including VLA,VLN,and Ego,and aligns physical…
  • It scores 0.
  • 620 overall on RBench,surpassing models such as Wan2.

Quick Facts

Published
Updated
Source
IT之家·人工智能
Category
Model Releases
AI score
72
Read the original Read the full article at the original source
Why it matters

蚂蚁灵波开源了首个面向具身智能的视频基模,用MoE控制推理成本,对机器人仿真和世界模型研究是个真工具,做具身的可以跑起来了。

Source & copyright

This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.

Original article · AIHOT entry

How to Cite

When citing this article, keep the title, date, and permalink, and credit Orange News as the source.

Orange News: Ant Lingbo open-sources LingBot-Video, the world's first MoE video foundation model for embodied intelligence. July 9, 2026. https://new.nat6.net/en/news/cmrd6bmv200xrih4b7i1ivun0/

View all →