Google DeepMind Releases Gemini Robotics-ER 1.6 — Advanced Spatial Reasoning for Real-World Robot Assistance

Google DeepMind released Gemini Robotics-ER 1.6 on April 15, a reasoning-focused AI model that substantially advances how robots perceive and interact with physical environments. The model introduces improved multi-camera perception, spatial awareness, and industrial instrument interpretation — enabling robots to identify objects, understand complex scenes, and interpret readings from gauges, screens, and controls with greater accuracy. A key development is improved safety performance: the model shows stronger hazard detection and avoidance behavior in unstructured real-world settings. For accessibility applications, improved spatial reasoning and visual interpretation capabilities in assistive robotics could enhance navigation aid for blind users, object retrieval assistance for mobility-impaired individuals, and autonomous assistance in healthcare settings where robot-assisted care is being explored. The release continues DeepMind's strategy of iterative improvements to reasoning capabilities that directly benefit physical AI systems in real-world deployment.