Alibaba· QwenNew model

Qwen-RobotNav: A Scalable Navigation Model Designed for an Agentic Navigation System

Alibaba presents Qwen-RobotNav, a scalable navigation model built on Qwen3-VL with a parameterised interface for controllable observation parameters. Trained on 15.6 million samples, it achieves state-of-the-art across five navigation domains and is designed as a reconfigurable primitive for agentic navigation systems, with consistent scaling from 2B to 8B parameters.

From the source

We present Qwen-RobotNav , a scalable navigation model built on Qwen3-VL that addresses this through a parameterised interface with two complementary dimensions: task modes that select the navigation behaviour, and controllable observation parameters (token budget, temporal decay, per-camera weights) that govern how visual history is encoded.

qwen.ai

Announced

16 Jun 2026 00:00 UTC

Detected