Google has released Gemini Robotics Embodied Reasoning 2 (ER 2) to Google Cloud, giving developers more options for developing physical AI.
Google released Gemini Robotics 2 earlier this year, but it has now released a public preview of ER 2 to Google Cloud. The preview is designed to help developers tap into the “enterprise-grade security, multi-agent orchestration, and end-to-end lifecycle management of Google Cloud — making it easier to simulate, evaluate, and deploy safe, adaptive physical AI at scale.”
AI development has been proceeding at a breakneck pace, but many researchers and companies are focused on bridging the gap between AI and physical robots. Once the two can be successfully married, it will open up a whole new—and potentially terrifying—world of possibilities.
Google emphasizes the deep reasoning necessary to make a success of this transition, reasoning that ER 2 brings to the table.
In physical operations, a robot needs to know whether a step actually succeeded before moving on to the next one. Gemini Robotics ER 2 adapts to new and unfamiliar situations on the fly by reasoning over continuous video feeds. The model:
- Classifies task progress into binary status, discrete categories, or continuous scores to improve real-time situational awareness and enable real-time adjustment or iteration on failed steps, without restarting an entire workflow.
- Pinpoints the exact video frame where a critical event occurs (such as when to stop pouring a liquid) with 91.3% accuracy and a 0.96-second mean absolute frame distance, running at 4x the execution speed and a fraction of the compute cost of many larger model categories.
- It advances our core spatial reasoning capability, including visual question answering and instrument reading. For example, it can read and interpret values across 10+ instrument categories including circular dials, sight glasses, digital displays, linear scales, rulers, and liquid thermometers directly from visual input.
Google has also created ER 2 to successfully scale, regardless of the size and intent of the robot:
Gemini Robotics ER 2 orchestrates robots of many shapes and sizes, scaling its intelligence from stationary arms and wheeled rovers to quadrupeds and complex forms. It also unlocks multi-robot collaboration, allowing different types of robots to share a semantic understanding of their workspace, divide labor, and hand off tasks to complete workflows a single machine could not do alone.
As with most frontier AI models, users will be able to interact with ER 2 using natural language:
Gemini Robotics ER 2 thinks about what comes next while simultaneously orchestrating actions. Developers can declare low-level control interfaces — such as VLA models, navigation APIs, or custom functions — as tools across simulation, real-world robot control, or human tele-operation. Gemini Robotics ER 2 can orchestrate navigation and manipulator movement APIs from natural language commands, allowing the robot to explain its approach while performing an action, and letting operators redirect it without using technical language.
Companies are already experiencing the benefits of ER 2.
“We’ve seen exponential growth in customers using AIVI-Learning every month, and Gemini Robotics ER 2 brings them even more value,” said Marco da Silva, VP and General Manager of Spot at Boston Dynamics. “Video changes what an inspection can be. Spot isn’t just checking whether something looks right in a moment in time, the ability to apply AI models to video opens a whole new world of possibilities.”