# Robostral Navigate: robot foundation model

Robostral Navigate is Mistral AI's first model for embodied navigation. An 8B parameter model that takes RGB images and natural language instructions to guide robots through complex environments. State-of-the-art on R2R-CE (76.6% success on validation unseen, 79.4% on validation seen). Operates with a single RGB camera, no LiDAR or depth sensors. Built entirely in-house, trained in simulation on ~400K trajectories across 6,000 scenes. Generalizes across wheeled, legged, and flying robots.

- **Slug:** robostral-navigate
- **Type:** Foundation model
- **Maturity (DEPLOY ladder):** Research
- **Open source:** no

## Architecture

8B parameter VLM initialized from vision-language model for grounding (pointing, counting, object localization). Navigation via pointing (infers image coordinates of target location) with fallback to metric displacements. Trained with prefix-caching (tree-based attention masking, 22x token reduction) and online RL (CISPO algorithm, +3.2% improvement).


## Key facts

- **Type:** Foundation model (embodied navigation)
- **Size:** 8B parameters
- **Input:** RGB images + natural language instructions
- **Sensors required:** Single RGB camera only; no LiDAR or depth sensors
- **Benchmark:** State-of-the-art on R2R-CE: 76.6% success on validation unseen, 79.4% on validation seen
- **Margin:** Beats best single-camera approach by 9.7 points; beats best multi-sensor by 4.5 points despite no depth sensors
- **Training:** Built entirely in-house; trained in simulation on ~400K trajectories across 6,000 scenes; no reliance on existing open-source VLMs
- **Efficiency:** Prefix-caching with tree-based attention masking compresses episode into single sequence; 22x token reduction vs per-timestep training
- **RL post-training:** Online reinforcement learning (CISPO algorithm) improved success rate by 3.2% after supervised training
- **Generalization:** Runs on wheeled, legged, and flying robots; generalizes across robot sizes and camera intrinsics
- **Navigation method:** Pointing-based (infers target coordinates in camera view) with fallback to metric displacements when target out of view
- **Authors:** Theo Cachet, Arjun Majumdar, Srijan Mishra, Thomas Chabal, Chris Bamford, Elliot Chane-Sane, Benjamin Tibi, Ludovic Ho Fuh, Olivier Duchenne (AI Science Robotics)


## Developed by (1)

- [Mistral AI](/companies/mistral-ai.md)


## Sources (3)

1. **Robostral Navigate: single-camera AI navigation (Mistral AI official)** · https://mistral.ai/news/robostral-navigate/ · 2026-07-08
2. **Mistral AI on X: Announcing Robostral Navigate, 8B robotics navigation model, SOTA on R2R-CE** · https://x.com/MistralAI/status/2075278815417528448
3. **Mistral AI robotics team hiring page** · https://jobs.lever.co/mistral?team=Research


## Common questions

### What is Robostral Navigate?

Robostral Navigate is Mistral AI's first model for embodied navigation. An 8B parameter model that takes RGB images and natural language instructions to guide robots through complex environments. State-of-the-art on R2R-CE (76.6% success on validation unseen, 79.4% on validation seen). Operates with a single RGB camera, no LiDAR or depth sensors. Built entirely in-house, trained in simulation on ~400K trajectories across 6,000 scenes. Generalizes across wheeled, legged, and flying robots.

### Who developed Robostral Navigate?

Robostral Navigate is credited to Mistral AI on the DEPLOY registry. Each developer attribution is verified via primary sources.

### Is Robostral Navigate open source?

No. Robostral Navigate is recorded as proprietary on the DEPLOY registry. Model weights and source are not publicly available.

### What type of AI is Robostral Navigate?

Robostral Navigate is a foundation model, built on a 8B parameter VLM initialized from vision-language model for grounding (pointing, counting, object localization). Navigation via pointing (infers image coordinates of target location) with fallback to metric displacements. Trained with prefix-caching (tree-based attention masking, 22x token reduction) and online RL (CISPO algorithm, +3.2% improvement). architecture on the DEPLOY registry.

### What is Robostral Navigate's maturity stage?

Robostral Navigate is at the research stage on the DEPLOY maturity ladder. Research stage means active development without commercial deployments on file.

### Which robots run on Robostral Navigate?

No robot models on the DEPLOY registry are recorded as running Robostral Navigate. DEPLOY wires brain-to-model connections only when the wiring is verifiable from primary sources; absence may reflect pre-deployment or unverified manufacturer claims.


_API: GET /v1/brains/75e9dffc-773c-4ef7-b0a0-1de2a0f9a850 · canonical URL: /brains/robostral-navigate_