Light Origins Open-Sources LightNav-0 Generalist Navigation Brain

Light Origins has open-sourced LightNav-0, a compact generalist navigation model built on Qwen3-VL-4B. Weights, code, a technical report and the INSIGHT-Bench evaluation set are published on GitHub and Hugging Face under Apache 2.0. The model uses a single token interface for instruction following, object-goal navigation and visual tracking.

• Built on Qwen3-VL-4B with no separate heads for different tasks
• Data engine turned 2,000+ real scenes into simulators and synthesized 4,000+ hours of data
• Three-stage training: mid-training, supervised fine-tuning and online RL
• Reported leadership among monocular methods in ten navigation settings