Google DeepMind announces Gemini Robotics 2

Adds new AI models to bring ‘whole-body intelligence’ to robots

Google DeepMind

By Robotics 24/7 Staff August 8, 2026

Google DeepMind

Google DeepMind said that Gemini Robotics 2 brings whole body intelligence to robots.

Email Sign Up

Get news, papers, media and research delivered. Sign up for our free newsletters.

Stay up-to-date with news and resources you need to do your job. Research industry trends, compare companies and get weekly market intelligence with Robotics 24/7.

Google DeepMind said that it demonstrated how Gemini's multimodal understanding could drive real-world action with Gemini Robotics.

Now, the company is introducing Gemini Robotics 2, which it calls the intelligence layer powering the next generation of truly adaptable robots. As it takes its first literal steps, DeepMind said that this major advance unlocks intelligent whole-body control, advanced dexterity and multi-robot collaboration.

Gemini Robotics 2 details

The company said that Gemini Robotics 2 enables robots to reason through every movement, unlocking a broad range of tasks. For example, DeepMind said that it can enable a humanoid to walk, crouch, stretch and manipulate objects to clean up a cluttered room. Additionally, the company said that it can even team up with other robots to finish the job faster.

DeepMind said that the company is making this possible through three highly capable models:

Gemini Robotics 2: Its most advanced vision-language-action model (VLA) that converts vision and language input into motor control, enabling a robot to take action. The company said that this model is capable of controlling full humanoids, from feet to fingertips, and other bi-arm robots. It also brings a new level of dexterous manipulation to both hands and grippers.
Gemini Robotics ER 2: DeepMind said that this is its most capable embodied reasoning (ER) model. It is a vision language model (VLM) that acts as its agent, enabling robots to communicate with humans, understand the physical world and plan multi-step tasks lasting several minutes. The company said that it is also introducing the ability for robots to work together as a team.
Gemini Robotics On-Device 2: The company’s most efficient vision-language-action model (VLA) is optimized to run locally on robotic devices. DeepMind said that this model can now achieve fast adaptation to completely new robot embodiments with a few hours of data.

Gemini Robotics ER 2, the reasoning model, is now available on Google AI Studio and in private preview on Gemini Enterprise Agent Platform. DeepMind added that VLA and On-Device models are available to early-access partners.

Click here to learn more details about Gemini Robotics 2, and how these models manage whole-body tasks for robots.

Latest in Mobile Manipulation

Latest in Artificial Intelligence

Latest Robotics News

Robotics 24/7

Google DeepMind announces Gemini Robotics 2