← All modelsFLUX 3 Action turns what the robot sees into what it does next. Give it the scene camera image, the wrist camera image, the current SO-101 joint state (shoulder pan, shoulder lift, elbow flex, wrist flex, wrist roll, gripper) and a plain-language instruction such as "Pick up the yellow cube and place it inside the black rectangle". It returns a chunk of 42 target joint positions at 30 Hz, that is 1.4 s of motion. In a control loop, execute the first 32 steps (about 1 s), then call again with fresh images and joint state.
Pricing
- Price
- 0.375 cr / second
Prices in credits (1 credit = $0.01).
Data schema
Input
| Field | Type | Description |
|---|
| seed | — | Sampling seed. Omit for a random seed; the reference evaluation seed is 0 for DROID and 42 for SO-101. |
| staterequired | array | Measured state: shoulder pan, shoulder lift, elbow flex, wrist flex, wrist roll (degrees), gripper (percent). |
| promptrequired | string | Natural-language task instruction for the robot. |
| state_history | — | Measured states at the 7 control ticks (30 Hz) before `state`, oldest first. Requires `command_history`. Omit for a fresh episode. |
| guidance_scale | number | Classifier-free guidance on video and action; 1.0 disables CFG. |
| command_history | — | Absolute commands in effect at each of the 8 observation ticks, oldest first; the last row is the most recent command sent. Requires `state_history`. |
| scene_image_urlrequired | string | Current frame from the fixed scene camera (tiled left). Resized to 256x256. |
| wrist_image_urlrequired | string | Current frame from the wrist camera (tiled right). Resized to 256x256. |
| num_inference_steps | integer | Euler denoising steps. The checkpoint was validated at 4. |
| history_scene_image_url | — | Scene frame from the oldest history tick (7 ticks before now). Requires `state_history`. Defaults to the current frame. |
| history_wrist_image_url | — | Wrist frame from the oldest history tick. Requires `state_history`. |
Output
| Field | Type | Description |
|---|
| seed | integer | Seed used for sampling. |
| actions | array | Predicted action chunk, one row per control step in time order. DROID: 32x8 absolute joint targets (rad) plus gripper closed fraction at 15 Hz. SO-101: 42x6 absolute commands (deg, gripper %) at 30 Hz. Not clipped: enforce joint, velocity and workspace limits on the robot. |
| timings | object | Server-side timings in seconds. |
| action_hz | number | Control rate of the action rows. |
| action_labels | array | Name of each action column. |
| execute_steps | integer | Execute this many leading actions, then request a new chunk. |
