← Leaderboard

1X World Model

World-action
1X Technologies · Jan 12, 2026

Video-pretrained world model used as NEO’s policy: it generates future frames from a text prompt and the current frame, and an inverse dynamics model turns them into actions. Results are self-reported on 1X’s own tasks (30 trials each; e.g. 30% success on tissue pulling with one generation, 45% with eight), and inference takes about 11 s per 5 s of video.

Type
World-action
Parameters
14B
Backbone
14B generative video model + inverse dynamics model
Weights
Closed
License
–
Commercial use
Unclear
Released
Jan 12, 2026
Robots
NEO
SO-101
–
LeRobot
–
Inputs
–
VRAM
–

Benchmark results

0 of 13 benchmarks

No benchmark results recorded yet.

Not reported on 13 benchmarks
Wrong or missing number? Report an error