Skip to content
AI.info

The Pulse

Google brings Hydrogen robotics tools to paid Antigravity users

Google has released a Robotics plugin for Antigravity, pairing robot setup, simulation and control tools with Gemini 4 Hydrogen. Antigravity access requires a paid Google AI Pro or Ultra subscription.

Google brings Hydrogen robotics tools to paid Antigravity users

Developers using Google Antigravity can now pair Gemini 4 Hydrogen with a new Robotics plugin to set up robot arms, build simulations and generate control applications. Antigravity is available to paid Google AI Pro or Ultra subscribers.

The plugin bundles tools for hardware onboarding, simulation and robot control, according to Google's November 17 developer post. Google introduced Hydrogen on November 12 as a model for coding and agentic workflows; the robotics release brings it together with the new Antigravity tools.

From instructions to robot control

Google says the model and plugin can take a developer from an unboxed robot to a working application in an afternoon. Given a plain-language goal and a camera view, Hydrogen plans the task, writes and runs control code, and adjusts its actions as it works.

On the Artificial Analysis Robotics 0.1 benchmark, Google reports that Hydrogen scored 58, similar to Gemini 4 Argon at a lower cost per token. The benchmark uses a real 6-axis arm for eight everyday tabletop tasks, including pouring a drink and packing blocks into a box.

On the public ERQA benchmark, Hydrogen outperforms Gemini Robotics ER 2 on spatial reasoning tasks such as matching views and reasoning about 3D trajectories, according to the company. Google recommends that developers using Gemini Robotics ER 2 try Hydrogen for those workloads.

Setup and simulation

The plugin packages skills for installing drivers, finding cameras, calibrating an arm and verifying motion. Its MuJoCo tools build and preview simulated scenes and assets inside Antigravity, where developers can adjust shapes, positions and physical properties through chat.

Developers can describe assets, import them from links or documents, or supply workspace photos. Another set of skills uses Gemini's spatial understanding to find objects, plan grasps and control a robot in a closed loop; application-building tools generate a controller and simulation.

In one Google demonstration, a developer connects an I2RT YAM arm and an external depth camera to a laptop. The company says the plugin installs software, finds the camera, calibrates the arm and runs test motions, leaving the arm ready for tasks in about 10 minutes. The post also describes setup flows for a SO-ARM101 arm and an FR3 Duo.

Applications and access

In a separate example, a developer starts with a sketch of a conveyor carrying mixed snacks for an arm to sort into bins. Google says Hydrogen constructs a MuJoCo simulation and tests the grasping policy before the developer deploys it to the physical robot. After deployment, the developer can request controller changes in natural language, including adjustments to gripper force or sorting rules.

The company reports that the arm was sorting snacks from the moving belt in about an hour from start to finish. Another demonstration connects a soap-shop web storefront to two YAM Pro arms: the arms set out a packing bin, pick the ordered bars and pull a printed thank-you note from a printer.

Users can select gemini-4-hydrogen in Antigravity and install the Robotics plugin through the platform's Customizations area. The model is also available through the Gemini API's Interactions API, with support in the google-genai SDKs; Google AI Studio supports spatial prompts, live camera streams and structured tool calls in the browser.

Google's model announcement sets pricing at $1.50 per million input tokens and $7.50 per million output tokens across the model's 1 million-token context window.

Sources

  1. Google Developers Blog
  2. Google

Companies in this story

Related stories

Explore

More articles