Gemini Robotics ER 2
An embodied reasoning model for robotics introduced by Google DeepMind, built on Gemini 3.6 Flash and optimized for robotic scenarios. It supports end-to-end reasoning capabilities encompassing perception, understanding, planning, and physical-world action execution.
Timeline (last 90 days)
Google has released Gemini Robotics ER 2, the latest generation embodied reasoning model for robotics based on the Gemini large language model.
Google's research team states that ER 2 demonstrates substantial improvements over the previous-generation ER model and the baseline model Gemini 3.6 Flash.
The official release tweet did not disclose detailed benchmark data for ER 2.
Specific performance metrics, the degree of openness, and commercialization plans for ER 2 have not yet been disclosed by the official team.
All Facts (3)
Google's research team states that ER 2 demonstrates substantial improvements over the previous-generation ER model and the baseline model Gemini 3.6 Flash.
50%UnverifiedThe official release tweet did not disclose detailed benchmark data for ER 2.
50%UnverifiedGoogle has released Gemini Robotics ER 2, the latest generation embodied reasoning model for robotics based on the Gemini large language model.
50%