Google DeepMind Launches Gemini Robotics ER2 Model

BlockBeats reports that on July 30, Google DeepMind launched the Gemini Robotics ER2 model, which is currently its most advanced embodied reasoning model. Its core positioning is as the "high-level brain" of robots, responsible for understanding the real physical world. It is capable of complex multi-step task planning and orchestration, progress tracking and failure detection, tool invocation (such as Google Search and custom functions), and directing the underlying Vision-Language-Action (VLA) model to execute specific actions. Together with Gemini Robotics 2 (VLA action model) and On-Device 2 (lightweight local version), it forms the complete Gemini Robotics 2 series. Click the original link below to join the Beating · Feishu AI News Channel for 24/7 global AI hotspot and news monitoring. [Original Link]