Google DeepMind uses Gemini 4 Argon to teach Apptronik’s Apollo humanoid new skills from video
Google DeepMind uses Gemini 4 Argon to teach Apptronik’s Apollo humanoid new skills from video
Google DeepMind has demonstrated a new robotics training technique where the Gemini 4 Argon AI model learns physical tasks by watching video and transfers the behavior to Apptronik’s Apollo humanoid. The model autonomously generates simulation code, enabling the robot to perform tasks like cornhole, achieving seven consecutive successful shots in real-world trials. The approach reduces simulation time from weeks to hours, showcasing the potential of general-purpose AI in robotics training.
Google DeepMind's Gemini 4 Argon model demonstrates a 'Video-to-Sim-to-Real' pipeline, enabling the Apollo humanoid to learn and execute physical tasks from video, reducing simulation time from weeks to hours.
Source: Robotics & Automation News · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
consecutive shots landed
7
View original evidence
ollo subsequently landed seven consecutive shots during a real-world trial. The company describes the approach as a “Video-to-Sim-to-Real” pipeline and says it can reduce a robotics simulation process that would normally require weeks of manual work to a few hours. Google DeepMind says: “This ‘Video-to-Sim-to-Real’ pipeline compresses weeks of manual simulation modeling into a few hours, transforming ordinary videos
Open source S4
- paper: not_reported
- code: not_reported
- weights: not_reported
- dataset: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
ollo subsequently landed seven consecutive shots during a real-world trial. The company describes the approach as a “Video-to-Sim-to-Real” pipeline and says it can reduce a robotics simulation process that would normally require weeks of manual work to a few hours. Google DeepMind says: “This ‘Video-to-Sim-to-Real’ pipeline compresses weeks of manual simulation modeling into a few hours, transforming ordinary videos
Open source S4
into a scalable, infinite training ground for robotics.” The demonstration is intended to show how general-purpose AI models could reduce the amount of manual engineering required to teach robots new physical skills. Google DeepMind also tested Argon on its Embodied Reasoning benchmark, or ERQA, which measures a model’s ability to reason about physical and spatial problems in 3D environments. The company says Argon a
Open source S5
chieved a score of 87 percent. Gemini 4 Argon is being positioned more broadly for complex engineering, scientific, coding and business tasks. Google says it has already used the model for projects ranging from optimizing large software codebases to quantum computing, nuclear fusion and life-sciences research. The model is now being made available through Google’s AI products and Gemini API, with Google presenting th
Open source S6
ode generation for robotics applications. In an internal test, Argon was shown raw video of a person playing cornhole. The model analyzed the environment, mapped its spatial coordinates and calculated the whole-body joint movements required to perform the task. Argon then autonomously generated simulation code before the learned behavior was transferred to Apptronik’s Apollo humanoid. According to Google DeepMind, Ap
Open source S3
Source：Robotics & Automation News · roboticsandautomationnews.com