Gemini Robotics-ER 1.6
#27 in Robotics & Embodied AIgoogle · v1.6 · since 2026-04-14 · 10× · last seen Jun 30, 2026
10
Momentum
Gemini Robotics-ER 1.6 is a Vision-Language Model (VLM) by Google DeepMind, based on the Gemini 3.0 Flash architecture, purpose-built for embodied reasoning in robotics applications. It acts as a high-level planning and reasoning layer for robots, enabling spatial logic, task planning, and success detection. The model processes image, video, and audio inputs alongside natural language commands, and can natively call tools such as Google Search and Vision-Language-Action (VLA) models. It is Google DeepMind's safest robotics model to date and represents a significant improvement over its predecessor ER 1.5.
Momentum trend
04.04.03.07.
Features
| Deployment Model | Cloud API (Hosted/Managed) via Gemini API and Google AI Studio; currently in Preview status; no on-premises or self-hosted operation documented |
| Platform | Gemini API, Google AI Studio (Preview); model name: gemini-robotics-er-1.6-preview; context window: 128k tokens (input), 64k tokens (output); based on Gemini 3.0 Flash; trained on Google TPUs |
| Release Date | April 14, 2026 |