Models · VLA (vision-language-action) · default

agi-smolvla/so101

checking availability…

Our checkpoint cagataydev/agi-smolvla-so101, for the SO-101 arm. It answers one observation with the next 50 actions. It is not a general-purpose controller; the task sentence is passed to the model, but only the task it was trained on is expected to work. The older id smolvla/so101 still works and leases the same row.

Price

per call: $0.005 · 50 tokens
per action: $0.0001
billed: every returned action, whether or not your robot plays it

What you send

observation.state [6]: shoulder_pan, shoulder_lift, elbow_flex, wrist_flex, wrist_roll, gripper · normalized: arm -100..100, gripper 0..100
observation.images.front [480, 640, 3]: front camera, RGB uint8 (JPEG on the wire)
observation.images.wrist [480, 640, 3]: wrist camera, RGB uint8 (JPEG on the wire)
task text: passed to the policy

Read the camera names out of your lease’s features (or the wire’s hello) rather than this table: a slot answers with its own.

What comes back

50 actions × 6 joints, to be played at 30 Hz — 1.67 s of motion per call.

Units before motion.Values are LeRobot normalized (arm −100…100, gripper 0…100), not degrees — whatever the semantics label says. That label echoes whichever runner answered your lease, so both values are in circulation: never branch on it (units, explained). LeRobot’s so101_follower defaults to degrees: run it with --robot.use_degrees=false. What the arm does if the connection drops.

Held out

ours: 14.45 chunk / 2.75 first step
the third-party checkpoint: 41.07 chunk / 28.33 first step

RMSE over its own 50-step chunk = 1.67 s of motion at 30 Hz / first step (normalised units) · lerobot/svla_so101_pickplace (apache-2.0), episodes 45–49 held out · fine-tune of lerobot/smolvla_base, 10k steps × batch 32, H200, 119 min. Lower is closer to the recorded human demonstration; it is not a task success rate.

The chunk here is its own 50 steps — 1.67 s of motion at 30 Hz. Do not read it against another model’s figure unless both the step count and the seconds of motion match: drift grows with time, so a chunk covering less of it is an easier average.

Try it

One billed call with synthetic inputs ($0.005). No robot moves.

Code

pip install agi-diy
agi login
agi run agi-smolvla/so101 --probe -n 1

Needs Python 3.12 — LeRobot is installed on first use, requires 3.12 or later and does not run on 3.14 yet.

Or run a test call from the dashboard, or drive the simulated SO-101.

Licences and data

Checkpoint	cagataydev/agi-smolvla-so101	apache-2.0
Base model	lerobot/smolvla_base,	no licence declared
Training data	lerobot/svla_so101_pickplace (apache-2.0), episodes 45–49 held out	as declared on the Hub

Licences as declared on the Hub. A lease says which of these answered it — pool and checkpoint on the session. Policies, explained.