Google DeepMind Releases Gemini Robotics-ER 1.6

Odaily Planet Daily News: Google DeepMind has released Gemini Robotics-ER 1.6, positioned as a high-level reasoning model for robots. Compared to its predecessors ER 1.5 and Gemini 3.0 Flash, it demonstrates significant improvements in spatial reasoning and multi-perspective understanding. The model is now accessible to developers via the Gemini API and Google AI Studio. Key upgrades include three capabilities: 1. Enhanced pointing accuracy: Enables precise object detection, counting, spatial relationship reasoning (e.g., "identify all objects that can fit into the blue cup"), and motion trajectory planning, while correctly rejecting pointing at objects that do not exist in the frame. 2. Successful multi-perspective detection: Robots can now integrate footage from multiple cameras to determine whether tasks are completed, maintaining accuracy even in occluded or dynamic environments. 3. Added instrument reading capability: Supports interpretation of various industrial instruments, such as circular pressure gauges, vertical liquid level indicators, and digital displays. Through agentic vision (visual reasoning + code execution), it achieves step-by-step reasoning by first zooming into detailed areas, then using pointing and code to calculate proportions and intervals, and finally combining world knowledge to derive readings.