title: "Robostral Navigate: navigazione AI a telecamera singola | Mistral AI"

Robostral Navigate is an 8B parameter model that enables robots to navigate complex environments using only a single RGB camera and natural language instructions. It achieves state-of-the-art performance on the R2R-CE benchmark, surpassing systems that rely on multiple sensors like LiDAR or depth cameras. The model was trained entirely in simulation using efficient techniques like prefix-caching and was further refined with online reinforcement learning to improve success rates. This approach, which combines visual pointing with local movement commands, allows the model to generalize across various robot types and adapt to real-world obstacles.