For robots to be really useful, they should perceive the bodily world like we do. That’s why right now we’re introducing Gemini Robotics-ER 1.6, an improve to our reasoning-first mannequin that allows robots to grasp their environments with unprecedented precision.
By enhancing spatial logic and multi-view understanding, we’re bringing a brand new degree of autonomy to the subsequent technology of bodily brokers. This mannequin focuses on capabilities crucial for robotics, together with visible and spatial understanding, job planning and success detection.
We’re additionally serving to robots with instrument studying, a brand new functionality to allow robots to learn advanced gauges and sight glasses — a functionality found via collaboration with Boston Dynamics. Gemini Robotics-ER 1.6 is our most secure robotics mannequin to this point, demonstrating superior compliance with security insurance policies on adversarial spatial reasoning duties.
Beginning right now, Gemini Robotics-ER 1.6 is obtainable to builders through the Gemini API and Google AI Studio. Learn extra on the Google DeepMind blog.
