Mistral enters robotics with Robostral Navigate, an 8B model that steers robots using just one camera
76.6% on R2R-CE. Mistral's new 8B robotics model navigates unknown spaces with a single camera—no lidar, no depth sensors.
Why it matters
Mistral is expanding beyond language into embodied AI, demonstrating that smaller, efficient models can handle real-world robot navigation. This signals a shift toward multi-modal, task-specific model releases and competition in the robotics AI stack.
The key facts
6 to knowMistral releases Robostral Navigate, an 8B model for robot navigation
76.6% benchmark score on R2R-CE (Room-to-Room Continuous Environment)
Single RGB camera input only—no lidar or depth sensors required
Trained in simulation with reinforcement learning (CISPO)
Availability timeline not yet announced
Marks Mistral's entry into embodied AI / robotics market
Go to the source
The Decoderthe-decoder.com
Publisher excerpt: Mistral is entering the robotics market with Robostral Navigate, an 8B model that guides robots through unknown environments using only a single RGB camera. Trained in simulation and refined with reinforcement learning (CISPO), it hits 76.6 percent on the R2R-CE benchmark. Mistral hasn't said when…