{"id":659654,"date":"2025-10-23T11:53:33","date_gmt":"2025-10-23T09:53:33","guid":{"rendered":"https:\/\/preprod.aivancity.ai\/blog\/deepmind-unveils-two-robotics-models-that-boost-robot-intelligence\/"},"modified":"2026-10-23T13:19:31","modified_gmt":"2026-10-23T11:19:31","slug":"deepmind-unveils-two-robotics-models-that-boost-robot-intelligence","status":"publish","type":"post","link":"https:\/\/aivancity.ai\/blog\/en\/deepmind-unveils-two-robotics-models-that-boost-robot-intelligence\/","title":{"rendered":"DeepMind unveils two “Robotics” models that boost robot intelligence"},"content":{"rendered":"\n

What if robots could finally think like us? Google DeepMind<\/strong> has just unveiled two artificial intelligence models, dubbed RT-X<\/strong> and AutoRT<\/strong>, capable of giving robots a much more nuanced understanding of their environment.These systems, developed through research in multimodal learning, mark a major step forward toward cognitive robotics<\/strong>, where machines no longer simply execute commands, but analyze, learn, and explain their actions<\/strong>.<\/p>\n\n

A Double Breakthrough: RT-X and AutoRT<\/h2>\n\n

DeepMind isn’t talking here about simple mechanical control models, but about true general-purpose intelligence architectures applied to robotics<\/strong>.<\/p>\n\n

Features<\/strong><\/strong><\/td>RT-X<\/strong><\/strong><\/td>AutoRT<\/strong><\/strong><\/td><\/tr><\/thead>
Model type<\/strong><\/td>A Unified Model of Robotic Reasoning and Action<\/td>Multi-robot orchestration and autonomy system<\/td><\/tr>
Main objective<\/strong><\/td>Understand and follow instructions in natural language<\/td>Plan, coordinate, and optimize the operations of multiple robots simultaneously<\/td><\/tr>
Apprenticeship<\/strong><\/td>Based on data from<\/strong> more than 30 laboratories<\/strong> and 17 billion parameters<\/strong><\/td>Continuous self-learning with autonomous feedback<\/td><\/tr>
Appetizers<\/strong><\/td>Vision, text, verbal interaction<\/td>Data from multiple sensors, vision, and performance feedback<\/td><\/tr>
Key Skills<\/strong><\/td>Contextual understanding, task transfer, explainable reasoning<\/td>Coordination, self-correction, robotic fleet management<\/td><\/tr>
Speed of adaptation<\/strong><\/td>Up to 60% faster<\/strong> than DeepMind’s previous systems<\/td>Real-time optimization powered by an automated scheduling engine<\/td><\/tr>
Areas of application<\/strong><\/td>Domestic, industrial, and experimental robotics<\/td>Multi-agent environments: warehouses, laboratories, hospitals<\/td><\/tr>
Battery life<\/strong><\/td>Language-based reasoning<\/td>Autonomous strategic control under human supervision<\/td><\/tr><\/tbody><\/table><\/figure>\n\n
\nhttps:\/\/youtu.be\/K4FYiK4CFp0?si=WhzzzgjDt2TFz9x4\n<\/div><\/figure>\n\n

This demonstration illustrates RT-X’s ability to interpret complex instructions such as “Sort the objects on the table by color”<\/em> and to adjust its movements autonomously.<\/p>\n\n

\nhttps:\/\/youtu.be\/a7BkJmTbohI?si=6f8BZi4_VG-Vt2h2\n<\/div><\/figure>\n\n

In this second video, DeepMind demonstrates how RT-X and AutoRT work together to manipulate objects, avoid obstacles, or coordinate multiple robots within the same workspace.<\/p>\n\n

An AI capable of understanding and explaining its actions<\/h2>\n\n

DeepMind’s models rely on a combination of computer vision, spatial reasoning, and natural language processing<\/strong>. Whereas older systems required specific training for each task, RT-X learns to generalize<\/strong>.<\/p>\n\n

By combining images, descriptions, and verbal instructions, it becomes capable of developing a comprehensive action plan and justifying its choices. A robot can thus explain why it chooses a particular route or decides to move one object rather than another.<\/p>\n\n

According to DeepMind, RT-X is based on a multimodal architecture with 17 billion parameters<\/strong>, capable of integrating visual and textual cues to understand the context of an action.<\/p>\n\n

Robots that learn like children<\/h2>\n\n

What sets these models apart is their ability to learn from their mistakes<\/strong>.AutoRT incorporates a self-evaluation mechanism that allows it to correct its actions and improve its performance without constant human supervision.<\/p>\n\n

DeepMind researchers compare this behavior to a formof developmental learning<\/strong>, similar to that of a child discovering the world through trial and error.<\/p>\n\n

This approach leads to robotics that is more autonomous, more adaptive, and capable of operating in unpredictable situations.<\/p>\n\n

A revolution in global robotics<\/h2>\n\n

The impact of these models goes beyond mere technical performance.DeepMind designed RT-X and AutoRT as open and collaborative<\/strong> systems: more than 30 international laboratories<\/strong> are participating in their development as part of the Open X-Embodiment<\/strong> project.<\/p>\n\n

This initiative aims to create a shared knowledge base<\/strong> among robots, where every learning experience can be shared.A robot trained in a Tokyo laboratory could thus instantly benefit from the experience of another robot based in Zurich.<\/p>\n\n

According to DeepMind\u2019s estimates, the combined use of AutoRT and RT-X could increase the speed at which robots adapt<\/strong> to complex environments such as warehouses, hospitals, or homes by 60%<\/strong> 1<\/a><\/sup>.<\/p>\n\n

Toward “thinking” robots<\/h2>\n\n

These advances mark a profound shift: the robots of the future will no longer be mere executors, but true reflective agents<\/strong>, capable of reasoning, planning, and explaining their decisions<\/strong>.<\/p>\n\n

An RT-X robot might say, \u201cI moved this object to avoid a collision<\/em>, \u201d<\/em> or \u201cThis surface seems unstable; I\u2019m choosing a different foothold.\u201d<\/em>This transparency, rare in robotics, marks a turning point toward explainable and responsible AI<\/strong> that is better integrated into human environments .<\/em><\/p>\n\n

Ethical challenges and the role of human oversight<\/h2>\n\n

The increasing autonomy of robots raises questions:<\/p>\n\n