Google DeepMind Launches Gemini Robotics ER 2 With Multi-Robot Collaboration

Google DeepMind has launched Gemini Robotics ER 2, its latest embodied reasoning model that helps robots plan complex tasks, monitor their progress in real time, and collaborate with other machines while adapting to changing conditions.



The model acts as a high-level physical agent, streaming video, audio, and text to orchestrate actions in real time while handing off motor execution to lower-level vision-language-action (VLA) models. It connects through the Gemini Live API using a bidirectional streaming endpoint, allowing robots to reason about upcoming steps while continuing to execute their current tasks instead of pausing to think between actions. Developers can also integrate tools such as Google Search or user-defined functions directly into the model's reasoning process.


A major upgrade is temporal intelligence, allowing robots to monitor their progress continuously and recognize when a task has actually been completed. Gemini Robotics ER 2 achieves 57.4% accuracy on continuous progress classification and 91.3% accuracy on moment finding, identifying the exact video frame where a critical event occurs with a mean absolute distance error of 0.96 seconds.

According to Google, that enables robots to transition between tasks, verify successful completion, and retry failed steps without restarting entire workflows.

Another major addition is multi-robot collaboration, allowing different robot types, such as a wheeled rover and a humanoid, to communicate through a shared semantic understanding and divide complex workflows that would be difficult for a single machine.

Safety has also been a focus. Gemini Robotics ER 2 halts a humanoid robot when a person is nearby and resumes work only once the area is clear. It also outperforms Gemini Robotics ER 1.6 and other frontier models on safety instruction following and human proximity benchmarks.


Gemini Robotics ER 2 is available now to developers through the Gemini API, Google AI Studio, and in private preview on the Gemini Enterprise Agent Platform.