Figure AI turned a warehouse-style task into a public test of human versus machine performance. Its humanoid robot F.03 competed against a human worker in a 10-hour livestreamed package sorting challenge. The goal was simple but physically demanding: identify the barcode on each parcel and place the box on the conveyor with the barcode facing down.

The human participant, Aime, narrowly won the contest. He sorted 12,924 packages, while Figure’s F.03 completed 12,732. The gap was only 192 packages, and the average speed was almost identical: 2.79 seconds per item for the human versus 2.83 seconds for the robot.

The comparison came with important context. Aime took legally required breaks and a meal period, while the robots were able to work continuously and rotate between units. By the end of the shift, the human reportedly suffered blisters and severe arm strain, while the machine continued operating after the contest.

Figure AI CEO Brett Adcock framed the result as the last time a human would win such a race. Even though the robot lost, the test highlighted how close humanoid systems are getting to human-level speed in repetitive logistics tasks. At the same time, the event also raised familiar questions about autonomy, reliability and whether humanoid robots can scale beyond controlled demonstrations into real industrial environments.