Unitree end-effector buyer guide Ā· Canada
A gripper maximizes repeatability. A dexterous hand maximizes research questions.
Dex1-1 and Dex3-1 are not good-better-best versions of one product. They use different mechanics, sensing and software interfaces for different object sets.
Quick answer
Choose Unitree Dex1-1 for repeatable parallel gripping, higher listed grasp loads, a 90 mm stroke and tasks with controlled object geometry. Choose Unitree Dex3-1 for seven-degree-of-freedom three-finger manipulation, tactile sensing and force-position research on a confirmed G1-compatible configuration. Start the quote with objects, poses and success criteria, not with the number of fingers.
Request an End-Effector Configuration ReviewTwo architectures
Parallel gripper versus articulated tactile hand.
Dex1-1 closes two long jaws across a large stroke. Dex3-1 coordinates a thumb, index finger and middle finger through seven active joints and tactile arrays.
Unitree Dex1-1
A 550 g parallel gripper with 90 mm stroke, adjustable 5 to 120 N gripping force and 0.1 mm fingertip position resolution. Unitree lists independent use plus compatibility paths for G1, H2 and R1.
Unitree Dex3-1
A 710 g three-finger hand with seven degrees of freedom, 33 pressure sensors and hybrid force-position control. Unitree publishes it as an option for G1 manipulation research.
Specification comparison
The strongest number is meaningful only inside its test condition.
Dex1-1's load figures and Dex3-1's 500 g grasp figure describe different geometries and test setups. They should not be ranked as if they were measured by one common test.
| Decision factor | Unitree Dex1-1 | Unitree Dex3-1 |
|---|---|---|
| Mechanism | Two-jaw parallel gripper | Three-finger articulated hand |
| Weight | 550 g | 710 g |
| Primary motion | 90 mm linear stroke | 7 active degrees of freedom |
| Sensing | Motor feedback including position, velocity, torque and temperature | Joint feedback plus 33 published pressure sensors and IMU data |
| Published load context | 20 kg form-fitting horizontal hold; 5 kg friction grip on a square steel block, both condition-dependent | 500 g on a 5 cm round hard object in two listed hand orientations |
| Communication | High-speed RS485, 6 Mbps, 1000 Hz listed control frequency | USB 2.0, 1000 Hz listed communication rate |
| Published platform fit | Independent use, G1, H2 and R1 | G1 option; exact robot and wrist configuration must be confirmed |
| Best research fit | Repeatable pick-and-place, force studies, fixture and sorting work | Tactile grasping, finger coordination, in-hand manipulation and imitation learning |
Robot arm payload is a separate constraint. The hand, adapter, wrist, cable, object and dynamic motion all consume payload. A hand's own grasp rating does not prove that the complete humanoid arm can safely move that load in every posture.
Task fit
Choose from the object set and research question.
Boxes, cylinders and repeatable parts
Prefer Dex1-1. Parallel surfaces, a defined stroke and adjustable force simplify fixturing, benchmarking and recovery.
Irregular objects and tactile contact
Prefer Dex3-1. Multiple fingers and pressure arrays create richer contact data, provided the object stays within tested load and geometry limits.
Sorting and cycle-time studies
Prefer Dex1-1 first. A simpler action space makes it easier to isolate detection, approach, grip and placement performance.
Imitation learning and grasp datasets
Prefer Dex3-1. Joint and tactile signals support more expressive datasets, but integration and calibration burden are higher.
Independent bench integration
Dex1-1 has the clearer published fit. Unitree explicitly lists standalone use. Confirm power, interface, mounting and software support in the quote.
G1 manipulation program
Either can fit a defined project. Confirm G1 model, EDU development access, wrist axes, hand quantity, compute, cables and software interfaces.
Procurement brief
A defensible quote names the complete manipulation system.
- List ten to twenty representative objects with size, mass, surface, compliance and fragility.
- Define the required grasp orientations, reach, wrist pose and placement tolerance.
- Set the cycle-time, success-rate and recovery targets.
- Name the robot model, EDU configuration, arm payload, wrist axes and development entitlement.
- Specify one or two end effectors, mounting hardware, cables, power and replacement parts.
- Define required feedback: position, force, tactile arrays, images, timestamps and logging rate.
- Plan a guarded bench test before free-space humanoid motion.
Review the live Unitree Dex1-1 and Unitree Dex3-1 paths, then browse the Unitree dexterous hands and grippers collection. Both are configuration-sensitive quote products rather than confirmed in-stock checkout items.
Unitree Dex1-1 vs Dex3-1 FAQ
What is the main difference between Unitree Dex1-1 and Dex3-1?
Dex1-1 is a two-jaw parallel gripper for repeatable force and stroke control. Dex3-1 is a seven-degree-of-freedom three-finger hand with tactile sensing for more complex manipulation research.
Which end effector has the higher listed grasp load?
Dex1-1 publishes higher load figures, but they depend on form-fitting or friction test geometry. Dex3-1's 500 g figure uses a different 5 cm round-object test, so the numbers are not directly comparable.
Does Dex3-1 include tactile sensing?
Yes. Unitree lists nine sensor arrays containing a total of 33 pressure sensors per Dex3-1 hand.
Can Dex1-1 be used without a humanoid robot?
Unitree lists independent use as a product compatibility path. The integrator must still confirm mounting, power, RS485 communication, software and safety controls.
Is Dex3-1 compatible with every Unitree humanoid?
No blanket compatibility should be assumed. Unitree publishes Dex3-1 as a G1 option. Confirm the exact robot, wrist, cable, controller and development configuration in writing.
Which is better for a first pick-and-place project?
Dex1-1 is usually the cleaner first platform because its simpler parallel motion reduces the control and dataset complexity of the grasp.
Which is better for tactile manipulation research?
Dex3-1 is the stronger fit because it provides articulated fingers, pressure arrays and richer joint feedback for contact-aware experiments.
Are Dex1-1 and Dex3-1 in stock at SpeedyDrone Canada?
The live pages use custom pricing and did not show ordinary in-stock checkout availability when checked on August 9, 2026. Confirm configuration, price and lead time with SpeedyDrone.
Official and live sources
- Unitree: Dex1-1 official product page - mechanics, force, stroke, load-test conditions, interface and compatibility.
- Unitree: Dex3-1 official product page - degrees of freedom, tactile sensing, load-test conditions, communication and G1 fit.
- SpeedyDrone: Unitree Dex1-1 - current Canadian quote path and real product images.
- SpeedyDrone: Unitree Dex3-1 - current Canadian quote path and real product images.
- SpeedyDrone: Unitree dexterous hands and grippers - current related product collection.
Quote the object set, robot and acceptance test together.
Send the target objects, robot configuration, task, payload, sensing, software and timeline. SpeedyDrone can help structure an end-effector configuration discussion before your team commits to hardware.
Request an End-Effector Configuration ReviewChecked August 9, 2026. Specifications, compatibility, software, included hardware, pricing and lead times can change. Manufacturer maximums are not guaranteed application results.
