Google DeepMind says Gemini 4 Argon outperforms frontier models on ERQA
🎧 Listen to this story
Google DeepMind says Gemini 4 Argon scored 87% on the Embodied Reasoning (ERQA) benchmark, which evaluates a model’s ability to reason through physical and spatial problems in 3D environments. The company says the result surpasses every other frontier model and that Argon can be used across different stages of robotics development.
The post presents examples in four areas: robot control, reconstruction of physically calibrated simulations, multimodal physical understanding, and improving actions through simulation and hardware iteration. Demonstrations include a Boston Dynamics Spot quadruped learning to use a foot pedal to dispose of cans, and a robotic arm reorienting a mug and placing a napkin inside it.
Why it matters · editorial interpretation
If Google DeepMind’s claimed result is confirmed, performance on ERQA may indicate that multimodal models can connect spatial reasoning with different stages of robotics. However, the score and demonstrations described do not by themselves establish reliability in real-world environments.
Sources
Research