
Gemini Robotics-ER 1.6
Embodied-reasoning brain for robot perception, planning, and tool use.
About
Gemini Robotics-ER 1.6 is Google DeepMind's embodied-reasoning model for robots, released April 14, 2026. The ER line is the high-level brain in DeepMind's robotics stack: it handles perception, spatial reasoning, task planning, and success detection, while lower-level vision-language-action models execute motor control. ER 1.6 improved instrument-reading accuracy from 23% on ER 1.5 to 86% baseline (93% with agentic vision enabled), reaching state of the art on 15 academic embodied-reasoning benchmarks, and is available to developers via the Gemini API and Google AI Studio.
Partners including Boston Dynamics, Apptronik, and Agility Robotics ship on the ER line, using it for facility inspections and manipulation planning. DeepMind describes ER 1.6 as its safest robotics model yet, with improved adherence to physical safety constraints such as refusing to lift objects beyond gripper limits. Benchmark figures are vendor-reported. The ER family has since continued with ER 2 (July 2026), which commands action models and APIs for multi-step work without stop-and-think pauses.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 92; Trending score 78/100 from researcher assessment.