An open-source 3B vision-language foundation model for on-device phone agents, with weights and training pipeline.