Google Gemini Robotics 2: AI Gets a Body — Here’s What It Can Do
Vídeo em Inglês, publicado pelo canal AI with Ain.
Assistir ao vídeo original no YouTube
Assuntos: Agentes de IA, Robótica e humanoides, AGI. Publicado no YouTube em 2026-10-04.
Seleção automática por assunto e idioma. O vídeo pertence ao canal de origem.
About Gemini Robotics
Gemini Robotics is a family of vision-language-action (VLA) models from Google DeepMind. A VLA model converts vision and language input into motor control, so a robot can act on what it sees and on the instructions it receives. According to Google DeepMind, the models are meant to power any type of robot, from bi-arm robots to full humanoids.
The current family has three models: Gemini Robotics 2, the most advanced VLA model; Gemini Robotics ER 2, an embodied reasoning model that plans multi-step tasks in physical spaces; and Gemini Robotics On-Device 2, a lightweight version that runs locally on robotic hardware. Google DeepMind lists whole-body control of humanoids, fine dexterity such as tying knots, and multi-robot collaboration among its capabilities. Access for testing is through a waitlist.
Fact sheet
- Name: Gemini Robotics
- Type: Vision-language-action (VLA) and embodied reasoning models for robots
- Developer: Google DeepMind
- Models: Gemini Robotics 2, Gemini Robotics ER 2, Gemini Robotics On-Device 2
- Access: Waitlist for testing
- Official page: deepmind.google/models/gemini-robotics
Read also
- Google's New AI Learned to Control a Robot From Just One Video
- The AI Frontier: Agentic AI, Reasoning, Robotics and the Path to AGI | AI Course Ep 23
- When AI Gets a Body: The Rise of Physical AI & Humanoid Robots 🤖⚡
Sources and references: Google DeepMind – Gemini Robotics · Wikidata (Q133568202) · video by AI with Ain on YouTube.