谷歌发布 Gemini Robotics 2：全身智能让机器人行走弯腰搬物
Original title: Google Introduces Gemini Robotics 2 with ‘Whole Body Intelligence’
Source: AI InsiderPublish time unverified
Google DeepMind 发布 Gemini Robotics 2 三模型家族：VLA 模型把视觉语言输入转为电机指令、ER 2 负责场景理解与多步规划、On-Device 2 可本地运行且少于 200 条样本适配新双臂机器人。新系统支持人形全身控制（行走、弯腰、货架放物）与多机器人协作，并配 ASIMOV-Agentic 基准测试拒绝危险动作的能力。
AI summary · AI generated
Insider Brief Google DeepMind has introduced Gemini Robotics 2, a family of AI models designed to give robots whole-body control, finer manipulation and the ability to work together on complex tasks. The release is aimed at moving robots beyond narrow, preprogrammed routines.