Figure 03 vs AgiBot X2-N

Figure 03 measures 173 cm x 50 cm x 45 cm and weighs 61 kg, featuring multi-density foam padding and soft textile covering for safe home interaction. AgiBot X2-N stands 131 cm tall, 46 cm wide and 21 cm deep, and a weight from about 35 kg. Both share similar lightweight humanoid forms but Figure 03 emphasizes compact, home-safe materials while AgiBot prioritizes proportions for dynamic tasks.

Figure 03 achieves 4.3 km/h speed using vision-language-action AI, SLAM, and mmWave data offload for adapting to changing home layouts. AgiBot X2-N walks at up to about 6.5 km/h via proprioceptive feedback and real-time gait adjustment, lacking GPS or LiDAR for self-reliant movement in dynamic settings. AgiBot offers superior speed, while Figure 03 excels in vision-based navigation for cluttered spaces.

Figure 03 integrates cameras including palm-mounted ones, tactile fingertip sensors detecting forces as low as three grams, and mmWave for fleet learning. AgiBot X2-N employs only proprioceptive sensors like joint torque, pressure, and gyros, with no cameras or external sensors. This results in Figure 03 having richer external perception versus AgiBot's internal feedback focus.

Figure 03 uses the Helix AI system, built on ROS and Linux, for pixel-to-action learning and voice commands. AgiBot X2-N features AI autonomous control with real-time systems on Linux/ROS, emphasizing proprioceptive-driven decisions. Both leverage ROS ecosystems, but Figure 03 incorporates vision-language models for versatile home autonomy.

Both robots specify a 3-5 year battery lifespan, aligning on long-term power design. Figure 03 includes 2.3 kWh capacity with 5-hour runtime and 2-hour charging, plus wireless options. AgiBot X2-N matches the lifespan without detailed runtime specs, suggesting comparable endurance for extended operations.

Figure 03 targets home assistance, chores, elderly care, and light manipulation like cleaning tasks in unpredictable layouts. AgiBot X2-N suits logistics, search-and-rescue, industrial automation, and dynamic navigation without visual reliance. Figure 03 fits domestic settings, while AgiBot excels in rugged, high-mobility industrial scenarios.

Figure 03 incorporates multi-density foam, soft textiles, emergency stop, obstacle avoidance, and pinch protection for home safety. AgiBot X2-N provides emergency stop, proprioception-based obstacle avoidance, and real-time gait adjustments. Both address core safety, with Figure 03 adding contact-soft materials.

Figure 03 runs Helix on ROS and Linux with Python/C++ APIs, voice/app control, and mmWave fleet learning. AgiBot X2-N uses Linux/ROS-based software for real-time AI control. Shared ROS foundation enables similar development, enhanced by Figure 03's vision integration.