Stone to Silicon

Artificial intelligence · 2025

Gemini Robotics

A frontier multimodal model folds origami and packs lunch boxes with robot arms.

Google DeepMind's Gemini Robotics (March 2025) added physical actions as an output modality to Gemini 2.0, generalising to unseen objects and robot bodies. With Physical Intelligence's pi0 and Figure's Helix, it marked robotics foundation models reaching dexterous general-purpose manipulation.

See it in the living treeWikipedia

Arcs above join it to what it stood on, arcs below to what it led to. The bar places this stretch on the tree's whole timeline.

What it built on 2

What it led to 4

Watch

Gemini Robotics: Bringing AI to the physical worldGoogle DeepMind · YouTube

Related measurements

The dashed line marks when it first worked.

Training compute of notable AI models Source: Epoch AI, Notable AI Models (CC BY 4.0)

ML hardware FLOP/s per dollar Source: Epoch AI, ML hardware (CC BY 4.0)