Alibaba QwenChina2026

Qwen-RobotNav

Runs onGo2by Unitree Robotics B

Key figures as stated by the developer

2 of 7 key figures stated

Inputs ⁠A camera language
Base model ⁠B Qwen3-VL

Not stated: Model type · Controls · Access · Parameters · Licence

Full specifications

6 of 14 fields stated

The letter points to the source at the foot of the page. Fields without a figure are dashed: the developer does not state them — that is not the same as zero, and not the same as no.
Field	Value	Source and caveats
Release
Released	16 June 2026	[A] "Submitted on 16 Jun 2026" ⁠C
Input and output
Model type	not stated	Note The developer does not label it vision-language-action, world action model or embodied reasoning: README subtitle: "A Scalable Navigation Model Designed for an Agentic Navigation System". ⁠A
Inputs	`camera` `language`	README, real-world deployment: "The only visual input is the Go2's built-in low-resolution camera"; commands are given as "step-by-step natural-language commands". ⁠A
Controls	not stated	Note The model predicts waypoints for the robot's path, not motor commands: README "so diverse navigation tasks can be expressed as trajectory prediction under different prompts and observation configurations." It was deployed on a Unitree Go2 quadruped. ⁠A
Degrees of freedom controlled	not stated
Size and speed
Parameters	not stated	Note The developer gives sizes per version, not one total: "performance improves from 2B to 8B parameters". ⁠B
Control rate	5.1Hz	End-to-end rate on a Unitree Go2, including network transmission. The developer's sources disagree on the setup: the technical report says "the remote-server deployment still achieves lower average end-to-end latency than on-device inference, with 196 ms (5.1 Hz) versus 204 ms (4.9 Hz)"; the blog says "We deploy Qwen-RobotNav on a Unitree Go2 quadruped robot with on-device inference via NVIDIA Jetson Thor, achieving 196ms latency (5.1 Hz)". ⁠B
Where it runs and what it is built on
On robot or cloud	both	Technical report: "we explore both cloud-based and on-device inference for Qwen-RobotNav". ⁠B
Base model	Qwen3-VL	Technical report: "zero architectural modification to the Qwen3-VL backbone". ⁠B
Openness
Access	not stated	Note README: "There is currently no plan to release the model weights for Qwen-RobotManip or Qwen-RobotNav." ⁠A
Licence terms	not stated
Licence	not stated
The developer’s own claims
Learns new tasks	not stated
Works across robots	yes	Developer claim, technical report: "Since Qwen-RobotNav operates across fundamentally different physical platforms, it must distinguish between embodiments such as indoor mobile robots and autonomous vehicles." The real-world deployment shown is on a Unitree Go2. ⁠B Yes only if the developer itself says the model runs on different robot models, for example robots from two manufacturers or two of one manufacturer’s models. Several units of the same robot do not count.

How to read the figures

50HzA: The letter in the grey square after a figure points to the source the figure came from. Pages with several sources carry several letters.

What the data labels mean

Notes

The Qwen team describes Qwen-RobotNav as a scalable navigation model.
The Qwen-RobotNav technical report was submitted to arXiv on 16 June 2026.

Versions and notes in the key figures (6)

Model type

Note The developer does not label it vision-language-action, world action model or embodied reasoning: README subtitle: "A Scalable Navigation Model Designed for an Agentic Navigation System".

Controls

Note The model predicts waypoints for the robot's path, not motor commands: README "so diverse navigation tasks can be expressed as trajectory prediction under different prompts and observation configurations." It was deployed on a Unitree Go2 quadruped.

Inputs

README, real-world deployment: "The only visual input is the Go2's built-in low-resolution camera"; commands are given as "step-by-step natural-language commands".

Access

Note README: "There is currently no plan to release the model weights for Qwen-RobotManip or Qwen-RobotNav."

Parameters

Note The developer gives sizes per version, not one total: "performance improves from 2B to 8B parameters".

Base model

Technical report: "zero architectural modification to the Qwen3-VL backbone".

Sources

A github.com/QwenLM/Qwen-RobotNav · Retrieved 2026-10-19
B arxiv.org/html/2606.18112 · Retrieved 2026-10-19
C arxiv.org/abs/2606.18112 · Retrieved 2026-10-19

One line per source, not per figure: a source has one retrieval date. The letter beside a figure points to the line it came from.