Compare

Kiki vs Reachy 2

← Back to robots

Quick verdict

Reachy 2 leads on 7 of 10 comparable rows

Kiki leads on Data Freshness, Price, and Social Interaction. Reachy 2 leads on Readiness Score, Reality Score, Battery, and Countries (+3 more).

Who should choose which?

Kiki
Choose Kiki if…
  • Budget is a priority and Kiki lists lower ($1,499).
  • Your main goal is conversation, which Kiki targets directly.
  • Strengths like social interaction and perception match how you plan to use it at home.
  • You prefer a desktop form factor over the alternatives in this compare.
Reachy 2
Choose Reachy 2 if…
  • You want higher readiness (86/100 vs peers) for near-term home use.
  • Verified shipping and public proof matter more to you (reality score 84/100).
  • Your main goal is education, which Reachy 2 targets directly.
  • Strengths like manipulation and perception match how you plan to use it at home.
  • You prefer a humanoid form factor over the alternatives in this compare.
Performance
Readiness Score
58
86
Reality Score
54
84
Data Freshness
100%
85%
SpeedNot mobileUnknown
PayloadUnknown3 kg per arm
Specs
Height25 cm136–166 cm
Weight1.8 kg32–50 kg
Battery4–6 h8–10 h
FormDesktopHumanoid
Availability
Price$1,499$70,000+
Market StatusDiscontinuedLimited
AvailabilityDiscontinuedLimited
CountriesUnited StatesUnited States, France, Global (direct order)
Intelligence
Mobility
Unknown
68
Manipulation
Unknown
88
Perception
80
86
Autonomy
62
72
Social Interaction
92
70
Home Navigation
5
58
SensorsNose camera (facial recognition), 4× beamforming mics, 16 touch sensors, accelerometer2× IMX296 global-shutter stereo cameras, Orbbec Gemini 336 RGB-D, Luxonis OAK ToF, RPLIDAR S2, 2× lavalier mics, wheel IMUs
ConnectivityWi-Fi, Bluetooth, iOS/Android companion appEthernet, Wi-Fi (customer network for AI)
EcosystemZoetic AI companion app (iOS and Android); Reinforcement learning personality engine; Neural imprint personality backup (announced)ROS 2 Humble; Python SDK; Hugging Face LeRobot; Pollen-Vision perception library; VR teleoperation (Meta Quest 2/3); Docker simulation stack (Gazebo, MuJoCo)
Green cells mark the stronger value in each row across 2 robots.