Maxwell smart: new Chinese embodied model tops global AI ranking on physical tasks
A new embodied AI model known as Maxwell – developed by the Chinese Academy of Sciences’ Institute of Artificial Intelligence for Industries – has shot to first place on the Meta-World benchmark for robot physical tasks. Maxwell scored 91.9 – the highest ever recorded on the simulation benchmark. Meta-World was set up by researchers from Stanford University, UC Berkeley and other institutions to…
A Chinese AI model called Maxwell, developed by the Chinese Academy of Sciences’ Institute of Artificial Intelligence for Industries, has topped the global AI ranking for physical tasks. Maxwell scored 91.9 on the Meta-World benchmark, which evaluates how well robots perform everyday physical tasks such as grasping, carrying, opening doors and using drawers.
The institute’s release noted that Maxwell could perform over 200 tasks without additional fine-tuning. Maxwell surpassed American rivals, who scored lower on the benchmark. Other top performers included FabriVLA from Shenzhen-based Youibot with 90, and SureFlow from Kyungpook National University in South Korea with 88.3. Maxwell's success also highlights the trend of Chinese companies making strides in this field, such as DeepCybo's PhysBrain, which topped 14 out of 28 benchmarks, and Unitree Robotics' foundation model, which can complete over 60 tasks.
The achievement of Chinese companies demonstrates their growing capabilities in creating intelligent robot models that can operate in the real physical world.
Written by urgent.news from South China Morning Post's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
