← Leaderboard

GEN-1

VLA
Generalist AI · Apr 2, 2026

Model trained from scratch on over 500,000 hours of physical-interaction data from human wearables, with no robot data in pretraining. Results are self-reported (99% average success on tasks where fine-tuned GEN-0 averaged 64%, with about one hour of robot data per task, and about 2.8x faster box folding than GEN-0 and π0); early-access partners only.

Type
VLA
Parameters
—
Backbone
–
Weights
Closed
License
–
Commercial use
Unclear
Released
Apr 2, 2026
Robots
–
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