Gemini Robotics-ER 1.6
Google DeepMind · 2026-04-14
Gemini Robotics-ER 1.6 is Google's preview vision-language model for advanced embodied reasoning in robotics. Based on Gemini 3.0 Flash, it specializes in spatial and physical reasoning, multi-view understanding, task planning, success detection, and tool-using task orchestration, and is available through the Gemini API and Google AI Studio.
Benchmark scores
| Benchmark | Score |
|---|---|
| Blueprint-Bench 2 | 0.0 |