ARTIFICE DAILY AI + CYBER SECURITY BRIEFING

Artificial Intelligence

Gemini 4 Argon targets robotics tasks

The post introduced Gemini 4 Argon as a physical reasoning model for robotics, said it scored 87% on ERQA, and showed example uses in manipulation, simulation and planning.

deepmind.google

The post presented Gemini 4 Argon as a model for robotics, code generation and visual-semantic reasoning, and said it was meant to support work across the robotics stack.

It said Argon reached 87% on the Embodied Reasoning (ERQA) benchmark, and the post described that result as better than every other frontier model tested there.

One example showed Argon controlling a Boston Dynamics Spot quadruped to dispose of soda cans in a foot-pedal bin, after first making an articulated MuJoCo digital twin of the bin to rehearse the move.

Other examples in the post covered a robotic arm placing a mug upright and putting a napkin inside it, arranging Scrabble tiles in a Gemini Omni World Model workflow, reconstructing teleoperation scenes in MuJoCo, and annotating an Aloha 2 origami-folding episode.

Named in this story

Products and systems

Gemini 4 Argon
The model the post introduced for robotics and physical reasoning
Embodied Reasoning (ERQA) benchmark
The benchmark used to report Argon’s score
Boston Dynamics Spot
The quadruped robot Argon was shown controlling
MuJoCo
The simulation platform used for the bin digital twin and other examples
Gemini Omni World Model
The system used in the Scrabble-tile rearrangement example
mujoco_warp
The simulation environment named in the throwing example
Aloha 2
The bimanual robot in the origami-folding example
Apollo 2 humanoid
The humanoid named in the throwing-sketch example

How the source tells it

The register was promotional and forward-looking, leaning on benchmark supremacy claims and broad statements about future robotics progress.

  • hype Loaded claims of a step change and being ahead of every other frontier model
  • speculation as fact Future-facing claims about accelerating progress towards physical AGI and changing the whole robotics stack