Google DeepMind Launches Gemini Robotics 2, Integrates Three Models for Humanoid Control and Multi-Robot Coordination
Google DeepMind launched Gemini Robotics 2, a physical AI system that combines three distinct models to provide whole-body control and advanced dexterity for humanoid robots. The primary architecture converts visual inputs and text instructions into movement, while a separate embodied reasoning model handles complex planning and multi-robot task coordination.
DeepMind demonstrated the system using Apptronik Apollo 2 and Franka Duo robots performing kitchen chores and garage organization. Real-world performance metrics show whole-body task success rates ranging from 45.7% to 76.3%, with fine manipulation accuracy varying from 32% to 92% depending on the specific chore. The ER 2 model, built on Gemini 3.6 Flash, is now accessible through the Gemini API and Google AI Studio.
From the sources (25 posts)
@googledeepmindOne brain. For any robot. 🤖 We’re launching Gemini Robotics 2: our next-generation physical AI bringing full body intelligence to humanoids, advanced dexterity, multi-robot teamwork and more.
@googledeepmindThree new models power this breakthrough: 1️⃣ Gemini Robotics 2: A vision-language-action model controlling humanoids from feet to fingertips 2️⃣ Gemini Robotics ER 2: Capable of real-world video understanding and complex, multi-step planni
@googledeepmindGemini Robotics 2 moves physical AI beyond tabletop tasks, enabling intelligent whole-body control for humanoids. Watch @Apptronik's Apollo 2 process a single prompt to reach, bend, and pick up a watering can ↓
@googledeepmindTo be genuinely useful in our homes and workplaces, robots need finesse. Gemini Robotics 2 equips different hardware platforms with high dexterity. It can control a five-fingered hand to tie a knot or screw in a lightbulb - while also mana
@financialjuiceGoogle is introducing Gemini Robotics ER 2. $GOOGL
RT @GoogleDeepMind: One brain. For any robot. 🤖 We’re launching Gemini Robotics 2: our next-generation physical AI bringing full body inte…
@_philschmidBipup beep... 🤖 Say hello to Gemini Robotics ER 2! ER 2 not only outperforms competing frontier models on embodied reasoning, it thinks and plans next steps simultaneously while executing current actions - Live API support for bidirectiona
@stockmktnewzGoogle $GOOGL DeepMind just posted this: “One brain. For any robot. We’re launching Gemini Robotics 2: our next-generation physical AI bringing full body intelligence to humanoids, advanced dexterity, multi-robot teamwork and more.” https:
@wallstengineGOOGLE LAUNCHES GEMINI ROBOTICS 2 $GOOGL introduced three models aimed at providing a shared intelligence layer across different robot platforms. Gemini Robotics 2 is a vision-language-action model that controls
@stocksavvyshay$GOOGL DeepMind launched Gemini Robotics 2 as “one brain for any robot” adding full-body intelligence, dexterity and multi-robot teamwork. The model is designed to power everything from humanoids to industrial arms and mobile robots. http
@officiallogankIntroducing Gemini Robotics ER 2, our latest robotics embodied reasoning model based on Gemini. So much progress from our last ER model (and baseline Gemini 3.6 Flash). Very excited to see more robots do useful things with this model! htt
@polymarketmoneyBREAKING: Google DeepMind launches Gemini Robotics 2 as “one brain for any robot” powering humanoids, industrial arms and mobile robots.
@officiallogankGemini Robotics ER 2 is available in the Gemini API and Google AI Studio. The demos of this model in action are very cool : ) pay special attention to multi-robot collaboration, this was my wow moment.
@techmemeGoogle DeepMind releases Gemini Robotics 2, which combines several different AI models into a single system to control a range of robots, including humanoids (@willknight / Wired) (Visit Techmeme dot com for the link and full context!)
RT @OfficialLoganK: Introducing Gemini Robotics ER 2, our latest robotics embodied reasoning model based on Gemini. So much progress from…
@tftc21Google DeepMind unveils Gemini Robotics 2, advancing AI for adaptable robots with whole-body control, dexterity, and multi-robot collaboration.
@testingcatalogGoogle introduced Gemini Robotics ER 2, a new embodied reasoning model! Benchmarks 👀 > Success/failure detection: Now operates on raw video feeds rather than static snapshots to catch mid-execution failures like spills, slips, or misalign
Introducing Gemini Robotics 2: the next era of truly adaptable robots from @GoogleDeepMind 🤖 Powered by three new models, it brings whole-body control, fine dexterity, and teamwork to complex tasks — making robots more helpful in the real
@osansevieroWelcome to Gemini Robotics ER 2! 🤖 A new embodied reasoning model for robotics, available in the Gemini API It can be used to understand physical world, plan multi-step tasks, chat with humans- from multi-robot collaboration to advanced s
Introducing Gemini Robotics 2: the next era of truly adaptable robots from @GoogleDeepMind 🤖 Powered by three new models, it brings whole-body control, fine dexterity, and teamwork to complex tasks — making robots more helpful in the real
@therundownaiGoogle just announced Gemini Robotics 2, extending its physical AI models beyond tabletop manipulation. A robot tying a trash bag sounds routine, but might be a bigger breakthrough than it seems:
@lmthangRT @GoogleDeepMind: One brain. For any robot. 🤖 We’re launching Gemini Robotics 2: our next-generation physical AI bringing full body inte…
@kimmonismusReally cool: Google DeepMind says Gemini Robotics 2 can control a humanoid’s entire body, from walking and crouching to manipulating objects with five-fingered hands. The system combines three models: Gemini Robotics 2 converts vision and
@demishassabisGemini Robotics 2 is here, with our new suite of models, robots can now reason through every movement to manage tasks that weren’t possible before, like tying delicate knots - and even team up to solve complex workflows. Huge congrats to th
@xiao_tedNew Gemini Robotics 2 release! What stood out to me: - VLA scales to real time whole-body humanoid control - Strong dexterity, from bi-arm grippers to 22DoF Sharpa hands - On-device cross-embodied VLA - Gemini Robotics ER 2 model new SOTA