[research] · · 1 min read
Google DeepMind’s Gemini Robotics 2 brings whole-body control to humanoid robots
The updated model lets robots walk, crouch, and use five-fingered hands for tasks like tying trash bags and unscrewing lightbulbs.
By ByteBulletin Editors · Editorial Team
Google DeepMind has unveiled Gemini Robotics 2, a major update to its AI model for controlling humanoid robots. While the previous version was limited to upper-body movements, the new model enables “whole-body motions” — from feet to fingertips — allowing robots to walk, crouch, stretch, and manipulate objects more naturally.
In videos shared by DeepMind, Apptronik’s Apollo 2 robot bends over to pick up a watering can and retrieves specific items from a shelf. The company acknowledges that movement speed still needs improvement, but calls the update “an important step towards the skills needed to complete more complex, real-world tasks that require whole-body coordination.”
Dexterity also gets a boost: Gemini Robotics 2 now controls complex five-fingered hands, enabling tasks like sealing a Ziploc bag, tying a trash bag, or unscrewing a lightbulb.
Alongside the robotics controller, DeepMind updated Gemini Robotics ER (embodied reasoning), a vision-language model that helps robots analyze surroundings and execute multi-step tasks. Gemini Robotics ER 2 improves task persistence and "now understands when tasks begin and end." It also supports multi-robot collaboration — one video shows Apollo 2 instructing Google’s dual-arm robot to put tools in a bin while cleaning a garage.
Safety is another focus. DeepMind says Gemini Robotics ER 2 is its “safest robotics model to date,” with better detection of nearby humans and the ability to trigger safety stops if someone approaches too closely.
Finally, the Gemini Robotics On-Device Model — which runs locally without an internet connection — now adapts faster to new robot embodiments, including those with "drastically different shapes, sensors and degrees of freedom."
For developers and roboticists, this release signals that embodied AI is moving beyond static, constrained demos toward more fluid, real-world interaction. The combination of whole-body coordination, fine dexterity, and safety-aware reasoning makes Gemini Robotics 2 a meaningful step toward general-purpose robots that can work alongside people.
SHARE
RELATED
[research] ·
Google Warns of 'Vishing' Attacks Targeting Financial Firms with Extortion Demands
Hackers are using phone calls to trick employees at major investment firms into handing over credentials, then extorting them for millions.
[research] ·
New Research Predicts LLM Inference Latency at the Edge, Aiming for Smarter Offloading
A new arXiv paper proposes a method to forecast LLM inference latency before deployment, which could make edge-device offloading decisions far more reliable.
[research] ·
Google’s AI Leadership Shake-Up: Turmoil or a Strategic Pivot?
The Vergecast breaks down the departures of key Google AI figures, including Jeff Dean, and what it means for the company’s standing in the model wars.