China · Commercial · 
Summary
X2 is AgiBot's compact humanoid robot for interaction, entertainment, education and embodied-AI development, supporting humanoid gait, expressive movement, multimodal interaction; X2 Ultra adds configurable autonomous navigation.
Identity
Status
Commercial
Category
Humanoid
Type
Humanoid
Function
Aid users
Audience
Developers
Physical
HEIGHT
~131cm
WEIGHT
35-39kg
PAYLOAD
max 3kg specific posture, <=1kg full workspace
Performance
DOF
25-30 depending on variant
NECK DOF
0(APP)/1(Ultra)
ARM DOF
5/arm(APP)-7/arm(Ultra)
LEG DOF
6
WAIST DOF
3
BATTERY
~421 Wh
MAX SPEED
1.8 m/s (daily <=0.8 m/s)
Other
SENSORS
3D LiDAR, RGB-D, RGB cameras, mic array
DIMENSIONS
1310x460x210mm
Video
Capabilities
RESEARCH (11 Dec 2025) — Object manipulation: work validated through experiments on AgiBot X2, reporting stronger performance than prior baselines across large-space manipulation tasks.
Height — ranked across 105 Humanoid platforms · this robot ranks #86
Weight — ranked across 88 Humanoid platforms · this robot ranks #74
Payload — ranked across 69 Humanoid platforms · this robot ranks #56
Degrees of freedom — ranked across 89 Humanoid platforms · this robot ranks #61
Comparable systems — same status tier
Other systems from AgiBot
Context & Benchmark
Comparable systems: AGIBOT A2 (AgiBot), GR-3 (Fourier Intelligence), Moya (DroidUp), Cinnamon 1 (Donut Robotics), Mornine (AiMOGA Robotics), AKINCI 5 (AkinRobotics). Related systems: LUNA (LimX Dynamics), Ultra Magnus (Chengdu APLUX), Panda Robot (UBTECH), Walker S2 (UBTECH), NEO (1X Technologies), Elixis (Addverb).
Timeline
PRODUCT LAUNCH (05 Jan 2026) — 'AgiBot Unveils X2-N Humanoid Robot at CES 2026': compact humanoid designed for entertainment, research and education; 25 DoF and touch sensors for human interaction; stands 1.3m tall, weighs 35kg.
Research
PAPER (23 Apr 2026) — 'X2-N paper presents transformable wheel-legged humanoid based on the AgiBot X2 line': high-DoF transformable robot operating in both humanoid and wheel-legged forms; pairs hardware with a reinforcement-learning whole-body controller for locomotion/transformation/manipulation; extends AgiBot X2 research into dual-mode locomotion, stair climbing, skating-like motion, and package-delivery experiments.