LightNav-0 open-sourced: one VLM-based model navigates four robot types autonomously
rsasaki0109 · x · 2026-10-14
The Light Origins team released LightNav-0, a compact generalist embodied navigation model that elicits spatial intelligence from a pretrained VLM (Qwen3-VL) and aligns it with navigation — no task-specific prediction heads.
In the demo, humanoid, quadruped, wheeled, and aerial robots operate in an unseen park, each following a language-named target with no teleoperation, fully autonomous. Diverse tasks share one token interface via dual-channel pointing. The project is open-sourced on GitHub under Apache-2.0 with habitat sim, mujoco demo, and real-robot deployment code.
Related event: LightNav-0 Open-Sourced: Qwen3-VL-Powered Navigation for Four Robot Types(2 posts)→
More from Embodied
- The rise of the robot 'deployment company': frontier models plus vertical hardware — chris_j_paxton · 2026-10-14
- A robot just rang the Nasdaq closing bell for the first time ever — JasonMa2020 · 2026-10-14
- NYC bans horse carriages; 'NYC 2030' mockup shows robot horses pulling the carts — robleclerc · 2026-10-14
- Dev ports Wright CAD workflow to ChatGPT Desktop, wiring Codex to SteveCAD for 3D design — burhop · 2026-10-14
- RoboStrategy claims the first AI robot to ring the Nasdaq bell, $BOT draws attention — cixliv · 2026-10-14
- Where Winds Meet to add NVIDIA DLSS 5 with 3D-Guided Neural Rendering in late November — ctnzr · 2026-10-14