The missing infrastructure for spatial intelligence. Complete.
Nutron is built from the ground up for training and operating robotics AI in the physical world.
Sifting through multimodal data, tuning robotics AI and observability — solved.
data platform AI tuning & observability
We built Nutron to make sense of fleet-scale data and train AI for any physical asset. Pair your embodiment with Nutron through nutron-cli and get started.
Complete stack for Spatial Intelligence
tracing
Alerts
metrics
mission api
fleet manager
RMIND
pi0
SmolVLA
Gr00t
starvla
& more
Data platform
indexing
alignment
search
curation
data debugging
Workflow
001
Ingest & index
002
Search & curate
003
Tune & deploy
004
Orchestrate tasks
Open source is at the heart of Nutron ♥️
L2D remains one of the most downloaded multimodal datasets for self-driving on 🤗. Our roadmap includes open-sourcing datasets from novel embodiments and environments. Subscribe below to be kept in the loop.
open-source Dataset & developer tools
L2D
L2D
World's largest open-source multimodal dataset. 90+ terabytes, 1Mio episodes, 30 cities & 60 vehicles.
Models
![](data:image/svg+xml,<svg display="block" role="presentation" viewBox="0 0 668 214" xmlns="http://www.w3.org/2000/svg"><path d="M 425.442 0 L 390.962 0 L 364.794 74.289 L 334.45 161.783 C 332.093 168.562 330.915 171.214 330.915 180.053 L 330.915 189.188 L 316.181 189.188 L 316.181 180.053 C 316.181 171.214 315.002 168.562 312.645 161.783 L 282.379 74.514 L 256.131 0 L 221.652 0 L 150.338 202.45 L 150.338 213.648 L 171.556 213.648 L 189.531 162.372 L 287.957 162.372 L 305.933 213.648 L 340.866 213.648 L 358.842 162.372 L 457.268 162.372 L 475.244 213.648 L 496.756 213.648 L 496.756 202.45 Z M 197.782 139.386 L 227.841 52.159 C 230.197 45.086 231.082 42.434 231.082 33.594 L 231.082 24.46 L 245.816 24.46 L 245.816 33.594 C 245.816 42.434 246.995 45.086 249.352 51.865 L 279.705 139.385 L 197.782 139.385 Z M 367.093 139.386 L 397.151 52.159 C 399.508 45.086 400.393 42.434 400.393 33.594 L 400.393 24.46 L 415.127 24.46 L 415.127 33.594 C 415.127 42.434 416.306 45.086 418.663 51.865 L 449.015 139.385 L 367.093 139.385 Z M 603.341 117.117 L 603.341 111.951 C 603.341 108.311 604.334 104.741 606.213 101.624 L 659.846 12.671 L 659.846 0 L 639.27 0 L 583.485 92.523 C 580.666 97.198 579.176 102.554 579.177 108.013 L 579.177 117.117 L 551.429 117.117 L 551.429 0 L 527.265 0 L 527.265 213.648 L 551.429 213.648 L 551.429 140.102 L 602.191 140.102 L 646.534 213.648 L 668 213.648 L 668 202.45 L 623.369 128.427 C 619.137 121.408 611.538 117.116 603.341 117.117 Z M 142.336 0 L 114.316 79.632 C 110.872 89.42 111.3 100.153 115.512 109.635 L 115.79 110.261 L 101.792 117.106 L 43.297 12.774 C 38.874 4.885 30.536 0 21.492 0 L 0 0 L 0 24.46 L 16.346 24.46 C 19.964 24.459 23.299 26.414 25.069 29.569 L 90.439 146.185 L 70.618 202.45 L 70.618 213.604 L 92.307 213.604 L 163.094 12.672 L 163.094 0 L 142.335 0 Z" fill="rgb(0, 0, 0)" height="213.6478159895544px" id="mlUOg__yb" width="668.0000143938418px"/></svg>)
rmind zero
Efficient foundation model for the edge.
Provider
Yaak
Parameters
Coming soon!
Embodiments
class B vehicles
Wheel loader
Fork lift
Terminal truck
dual-arm manipulator
pi05
Flow-matching VLA
Provider
Physical Intelligence
Parameters
~3.3B
PaliGemma VLM + action expert
Embodiments
Single-arm
Bi-manual
Mobile
SmolVLA
Compact open VLA
Provider
Hugging Face (LeRobot)
Parameters
~450M
SmolVLM-2 + action expert
Embodiments
Single-arm
SO-100 / SO-101
GR00T N1
Humanoid model
Provider
NVIDIA
Parameters
2.2B
Eagle-2 VLM (1.34B) + DiT
Embodiments
Humanoid
Bi-manual arms
Alpamayo 1
Reasoning AV model
Provider
NVIDIA
Parameters
10B
Cosmos-Reason (8.2B) + expert (2.3B)
Embodiments
Vehicles
Autonomous driving
Just add intelligence
Transform your physical assets into a fleet of autonomous workers
