Google made Gemini Robotics ER 2 available in public preview on its cloud AI platform on October 27. The model, identified as gemini-robotics-er-2-preview, is built on Gemini 3.5 Flash and supports interpreting physical surroundings, understanding video, planning tasks, and coordinating tools. The cloud offering lists an input limit of 131,072 tokens and an output limit of 65,536 tokens, with availability through the global endpoint.
Google says the cloud integration adds enterprise security and workflow management around a model designed to coordinate robot actions and assess whether tasks succeeded. It can support robot arms, wheeled platforms, and humanoids, according to the company. This expands access to an existing model; the cloud service remains a preview, with potentially limited support.
Source · cloud.google.com ↗
