OpenFrontier:
General Navigation with Visual-Language Grounded Frontiers

Esteban Padilla-Cerdio^* · Boyang Sun^* · Marc Pollefeys · Hermann Blum

^* equal contribution

RSS 2026

ArXiv | Webpage

Official code for OpenFrontier, a training-free navigation framework that integrates vision-language models for zero-shot robot navigation.

🚧 Code coming soon

We are working on releasing the code.

Citation

If you find our work useful, please consider citing:

@inproceedings{openfrontier2026,
  title     = {OpenFrontier: General Navigation with Visual-Language Grounded Frontiers},
  author    = {Padilla-Cerdio, Esteban and Sun, Boyang and Pollefeys, Marc and Blum, Hermann},
  booktitle = {Robotics: Science and Systems (RSS)},
  year      = {2026}
}