RH5DG2
commercialHumanoid

RH5DG2

The RH5DG2 is a Humanoid humanoid robot developed by Beijing Inspire-Robots Technology Co., Ltd.. It features 350 kg, 13 degrees of freedom.

Weight

350 kg

DOF

13

Specifications

Physical

Weight

350 kg

Degrees of Freedom

13

Locomotion

bipedal

Related news

Gemini Robotics 2: Google DeepMind's Humanoid Robot AI Explained

Google DeepMind released Gemini Robotics 2 on July 30, 2026, its first vision-language-action model capable of controlling a full humanoid robot including legs, torso, arms, and multi-finger hands under a single learned policy. The model demonstrates success rates ranging from 92% for unscrewing a light bulb to 32% for sweeping, with the strategic intent of supplying AI intelligence to robot manufacturers rather than building hardware itself. Apptronik's Apollo 2 serves as the primary demonstration platform, while Boston Dynamics also signed a partnership to integrate Gemini Robotics into its next-generation Atlas.

blog.robozaps.com•Jul 31, 2026•via Beijing Inspire-Robots Technology Co., Ltd.

Gemini Robotics 2 Controls Full Humanoids: Legs, Torso, Arms, and Fingers Under One Policy

Google DeepMind released Gemini Robotics 2, the first vision-language-action (VLA) model capable of controlling a full humanoid body—including legs, torso, arms, and multi-finger hands—under a single unified policy, collapsing the traditional separation between locomotion and manipulation controllers. The release includes three models: the flagship VLA, the embodied reasoning model (ER 2) now available to developers via the Gemini API, and an on-device version that can adapt to new robot embodiments in under 200 demonstrations.

techtimes.com•Jul 30, 2026•via Beijing Inspire-Robots Technology Co., Ltd.

Gemini Robotics 2 brings whole body intelligence to robots

Google DeepMind announced Gemini Robotics 2, a major advancement in robotics AI that enables intelligent whole-body control, advanced dexterity, and multi-robot collaboration. The system includes three models: a vision-language-action model (VLA) for controlling humanoids, an embodied reasoning model (ER) for high-level planning, and an on-device model for local operation that can adapt to new robot embodiments in hours.

deepmind.google•Jul 30, 2026•via Beijing Inspire-Robots Technology Co., Ltd.

More from Beijing Inspire-Robots Technology Co., Ltd.