PulseAugur
EN
LIVE 07:46:01
中文(ZH) 蚂蚁灵波同日开源两大模型,���动世界模型与具身智能视频生成的双升级

Ant Group releases two open-source AI models for world generation and video synthesis

Ant Group's Lingbo Technology has released two new open-source models: LingBot-World 2.0 for real-time interactive world generation and LingBot-Video, an open-source foundation model for embodied AI video generation. LingBot-World 2.0 supports high-definition output and integrates an Agent mechanism, while LingBot-Video is notable for its MoE architecture and its application in robotics for understanding actions and the physical world. AI

IMPACT These releases provide new open-source tools for real-time world generation and embodied AI video synthesis, potentially accelerating research and development in robotics and virtual environments.

RANK_REASON Open-source release of new AI models by a major tech company. [lever_c_demoted from frontier_release: ic=1 ai=1.0]

Read on 36氪 (36Kr) →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Ant Group releases two open-source AI models for world generation and video synthesis

COVERAGE [1]

  1. 36氪 (36Kr) TIER_1 中文(ZH) ·

    Ant Group Open Sources Two Major Models on the Same Day, Driving Dual Upgrades in World Models and Embodied Intelligence Video Generation

    7月9日,蚂蚁灵波科技连续开源LingBot-World 2.0与LingBot-Video两款新模型,分别面向实时交互世界生成和具身智能视频生成。前者支持小时级实时生成、720p/60fps 高清输出,并首次将Agent机制引入世界模型;后者则是全球首个基于MoE架构、面向具身智能的开源视频生成基础模型,为机器人理解动作、任务和物理世界提供新的开源底座。