Google DeepMindUnited Kingdom2026

Gemini Robotics 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.

RobotsApollo 2by Apptronik A

Key figures as stated by the developer

4 of 6 stated · 2 gaps

Model type ⁠B vision-language-action
Controls ⁠A arms and hands whole body walking
Inputs ⁠A camera language
Access ⁠A partners-early access only
Parameters not stated
Base model not stated

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 (4)

Model type

Developer wording: "Our most advanced vision-language-action model (VLA) that converts vision and language input into motor control".

Controls

"This model is capable of controlling full humanoids, from feet to fingertips, and other bi-arm robots." "enable a humanoid to walk, crouch, stretch, and manipulate objects to clean up a cluttered room".

Inputs

Google DeepMind says the model "converts vision and language input into motor control".

Access

"Our VLA and On-Device models are available to early-access partners".

Full specifications

6 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`	B Developer wording: "Our most advanced vision-language-action model (VLA) that converts vision and language input into motor control".
Inputs	`camera` `language`	A Google DeepMind says the model "converts vision and language input into motor control".
Controls	`arms and hands` `whole body` `walking`	A "This model is capable of controlling full humanoids, from feet to fingertips, and other bi-arm robots." "enable a humanoid to walk, crouch, stretch, and manipulate objects to clean up a cluttered room".
Degrees of freedom controlled	not stated	Note No total is given for the model. The only count in the source belongs to a hand, not to the model: "five-fingered, 22 degree-of-freedom SharpaWave hand on the Apollo 2 robot". ⁠A
Size and speed
Parameters	not stated
Control rate	not stated
Where it runs and what it is built on
Runs where	not stated	Note Where the model runs is not stated for Gemini Robotics 2 itself. The local model is a separate one, "Gemini Robotics On-Device 2", which is "optimized to run locally on robotic hardware". ⁠B
Base model	not stated
Openness
Access	partners-early access only	A "Our VLA and On-Device models are available to early-access partners".
License	not stated
The developer’s own claims
Learns new tasks	without examples	B Developer claim: "Gemini Robotics 2 can complete a variety of tasks - even if it hasn't been trained on them before."
Works across robots	yes	A Developer claim: "Gemini Robotics 2 controlling three different embodiments, using the same model checkpoint": "the Apptronik Apollo 2 robot with SharpaWave hands, the Apollo 2 robot with Inspire hands, and the Franka Duo with the Robotiq gripper". 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 calls Gemini Robotics 2 its most advanced vision-language-action model; it was announced on 30 July 2026 with a separate reasoning model (Gemini Robotics ER 2) and an on-device model (Gemini Robotics On-Device 2).
The reasoning model, Gemini Robotics ER 2, is a vision language model that plans tasks and coordinates with the vision-language-action model; it is a separate model.
Google DeepMind published the announcement on 30 July 2026.

Sources

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