Unitree G1: full technical specifications, real capabilities and limits

Unitree G1: the main technical specifications

The G1 comes in several configurations. Specifications vary in particular between the standard G1 and the G1 EDU versions intended for development. The standard Unitree G1 has 23 degrees of freedom, while some EDU configurations go up to 43.

A degree of freedom is an independent movement that a joint can perform. The number gives an indication of how rich the robot's range of gestures is: walking, changes of posture, independent arm movements and, depending on the configuration, additional joints in the upper body or the hands.

Perception, battery life and power: what do the figures mean?

The G1 combines a 3D LiDAR and a depth camera. The LiDAR measures the distance of objects around the robot and produces a three-dimensional representation of its surroundings. The depth camera complements this perception. These sensors are an important technical foundation, but they do not mean that a humanoid can be dropped in the middle of a trade show and move around on its own all day.

Unitree quotes about 2 hours of battery life with a removable 9,000 mAh battery. This figure should be read as a manufacturer's indication: consumption depends heavily on the movements. A relatively static posture, occasional walking and a light sequence of gestures do not draw on the battery in the same way as a succession of dynamic movements.

At events, we therefore think in terms of entertainment sequences rather than continuous dynamic operation. This approach makes battery management easier and, above all, preserves the highlights: a humanoid draws more attention when it appears at chosen moments than when it stays active without a break.

Joint torque is the rotational force produced by the motors at the joints. Weighing about 35 kg and with joints capable of fast movements, the G1 can produce very dynamic gestures. The flip side of this capability is the need to keep a suitable movement area and not to treat the machine as a toy.

What the Unitree G1 can really do at an event

For an event, the G1's value lies mainly in its physical presence and its ability to produce movements that are immediately recognisable as humanoid. It can walk around a defined area, perform prepared choreographies, adopt different postures, pose with visitors, accompany a stage entrance or create a focal point in front of a stand.

• Walkabouts within a defined, suitable area.
• Prepared choreographies and gesture sequences.
• Photos and videos with visitors.
• Stage entrances, product launches or grand openings.
• One-off entertainment on a stand or at a corporate party.
• Carrying light props when the configuration and the sequence allow it.

In this context, the robot becomes a tool for event scripting. Two services using the same G1 can produce very different results: the choice of movements, their timing, the robot's position, the space available and the operators' work matter as much as the spec sheet. This is the approach we take in our humanoid robot services.

Having operators present remains essential in the middle of a crowd. It makes it possible to adapt the robot's movements to the density of visitors, to handle unexpected interactions and to interrupt a sequence if the surroundings are no longer suitable. The aim is not to demonstrate maximum theoretical autonomy, but to deliver controlled entertainment under the real conditions of the event.

What the Unitree G1 does not do

Demonstration videos of humanoids can give the impression that a robot like the G1 spontaneously understands its surroundings and can carry out any task. That is not the case. A G1 is not an autonomous artificial human able to arrive in an unfamiliar place, understand a complex task on its own and improvise all its interactions among hundreds of people.

Some demonstrations rely on specially developed movements, controlled environments, dedicated algorithms or prepared sequences. The standard G1 must also be distinguished from the EDU versions, which meet different development needs.

For an event, the useful rule is simple: you do not build a service around what the robot could theoretically do, but around what it can perform in a controlled way at the venue, in front of the audience and for the planned duration.

Buying the robot can make sense for a company, a laboratory or a technical team that wants to develop robotic applications over the long term. For a one-off operation, the robot is only one part of the set-up: transport, preparation, batteries, operation, sequences and management of the surroundings must also be taken into account. Hire with operators spares the organiser from having to run the system themselves.

The spec sheet is therefore not enough to choose a humanoid. The decisive questions are often very practical: space available, crowd density, type of floor, walkabout or fixed area, indoors or outdoors, length of the sequences and the effect you are after. The G1 is a mobile, expressive humanoid platform; its value at an event comes above all from the way its capabilities are scripted in front of a real audience.