Google DeepMind Launches Gemini Robotics 2 for Whole-Body Robot Intelligence

gabriberton · x · 2026-08-07

Google DeepMind has introduced Gemini Robotics 2, a next-generation physical AI model designed to power a new era of adaptable robots.

As an advanced Vision-Language-Action (VLA) model, it introduces three major capabilities:

Whole-body intelligence: Enables humanoids to autonomously perform complex movements like walking, crouching, stretching, and cleaning cluttered rooms.
Advanced dexterity & teamwork: Supports fine object manipulation and allows multiple robots to collaborate and finish tasks faster.
On-device & fast adaptation: Runs locally on-device and can seamlessly adapt to entirely new robotic bodies in just a few hours.

More from Embodied