Google DeepMindUnited Kingdom2026

Gemini Robotics On-Device 2

An AI model: a vision-language-action model (VLA) takes camera images and an instruction and turns them into movements for a robot. This page describes the software, not a robot.

Robotsnot stated

Key figures as stated by the developer

5 of 6 stated · 1 gap

Model type ⁠A vision-language-action
Controls ⁠A arms and hands
Inputs ⁠A camera language robot state
Access ⁠A partners-early access only
Parameters not stated
Base model ⁠A Gemma (on-device models)

How to read the figures

14kgA30.9lbA: The letter in the grey square after a figure points to the source the figure came from. Robots with several sources carry several letters.
74.5lb: Imperial figures are our own conversion unless the manufacturer states an imperial figure itself. The source letter always belongs to the metric figure — a conversion has no source of its own.

The four states

Versions and notes in the key figures (5)

Model type

Developer wording: "Gemini Robotics On-Device is a state-of-the-art Vision-Language-Action (VLA) model based on our on-device Gemma models.".

Controls

Model card, Outputs: "Robot actions as numerical values." The card says the model "is trained to provide a general base for a variety of tasks on bi-arm robots" and "risks associated with mobile platforms or whole-body control are outside its current scope".

Inputs

Model card, Inputs: "Text (e.g., a question or instruction), images (e.g., robot’s perspective of the environment), and robot proprioception as numerical values."

Access

"Gemini Robotics On-Device 2 is distributed only to Trusted Testers".

Base model

"Gemini Robotics On-Device is an on-device VLA model based on Gemini Robotics 1.5 technology and our on-device Gemma models."

Full specifications

7 of 12 fields stated

The letter points to the source at the foot of the page. Fields without a figure are dashed: the developer does not state them — that is not the same as zero, and not the same as no.
Field	Value	Source and caveats
Input and output
Model type	`vision-language-action`	A Developer wording: "Gemini Robotics On-Device is a state-of-the-art Vision-Language-Action (VLA) model based on our on-device Gemma models.".
Inputs	`camera` `language` `robot state`	A Model card, Inputs: "Text (e.g., a question or instruction), images (e.g., robot’s perspective of the environment), and robot proprioception as numerical values."
Controls	`arms and hands`	A Model card, Outputs: "Robot actions as numerical values." The card says the model "is trained to provide a general base for a variety of tasks on bi-arm robots" and "risks associated with mobile platforms or whole-body control are outside its current scope".
Degrees of freedom controlled	not stated	Note The card says the model is "limited in its ability to generalise to out of distribution tasks as well as controlling high-degree-of-freedom robots". ⁠A
Size and speed
Parameters	not stated
Control rate	not stated
Where it runs and what it is built on
Runs where	on the robot	B "optimized to run locally on robotic devices".
Base model	Gemma (on-device models)	A "Gemini Robotics On-Device is an on-device VLA model based on Gemini Robotics 1.5 technology and our on-device Gemma models."
Openness
Access	partners-early access only	A "Gemini Robotics On-Device 2 is distributed only to Trusted Testers".
License	not stated
The developer’s own claims
Learns new tasks	not stated	Note The adaptation Google DeepMind describes is to new robot bodies: "We can now adapt to new bi-arm robot embodiments with just a few hours of adaptation time, typically with less than 200 examples." ⁠B
Works across robots	yes	B Developer claim: "This model is natively multi-embodiment"; "as shown below with a diverse set of tasks being performed by the Dexmate, SO101, and Trossen platforms". Yes only if the developer itself says the model runs on different robot models, for example robots from two manufacturers or two of one manufacturer’s models. Several units of the same robot do not count.

Notes

Google DeepMind describes Gemini Robotics On-Device 2 as a lightweight version of its vision-language-action model, optimized to run locally on robotic hardware.
Google DeepMind published the model card on 30 July 2026.

Sources

A Manufacturer product page. deepmind.google/models/model-cards/gemini-robotics-on-device-2/ · Retrieved 2026-10-13
B deepmind.google/blog/gemini-robotics-2-brings-whole-body-intelligence-to-robots/ · Retrieved 2026-10-13

One line per source, not per figure: a source has one retrieval date. The letter beside a figure points to the line it came from.