Rowe pricing is based on input and output spatial tokens. Input pricing includes the base cost of running the spatial reasoning job, while output pricing tracks the structured result Rowe produces. Billing is disabled during developer preview, but usage is still tracked in the dashboard so you can see what production would cost.
| Model | Input spatial tokens | Output spatial tokens | Preview | Best for |
|---|---|---|---|---|
Rowe 1.0 Developer preview | $0.06 / 1M | $0.12 / 1M | Billing off | Spatial reasoning, prediction, action planning |
Billing is currently off while Rowe access is opened to early developers. The dashboard still shows API calls, input/output spatial token estimates, and projected cost so teams can plan before billing turns on.
Input spatial tokens measure the scene context you send: point clouds, images, sensors, goals, object ids, and temporal history.
Output spatial tokens measure Rowe's structured result: predictions, reconstructions, action plans, fit checks, and explanations.
Requests are served through the Four Echelon API gateway and routed to available Rowe workers.
Pricing follows the request shape you already use: send Rowe a spatial scene and read the usage object returned with the response.
curl https://fourechelon.com/api/v1/responses \
-H "Authorization: Bearer $FOUR_ECHELON_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "rowe-1.0",
"input": {
"task_type": "predict",
"point_cloud_scene": {
"point_clouds": [
{
"object_id": "part-a",
"points": [
{ "x": 0, "y": 0, "z": 0 },
{ "x": 1, "y": 0, "z": 0 },
{ "x": 0, "y": 1, "z": 1 }
]
}
]
}
}
}'Create a key, test the playground, and track usage from one place.