GEN-1

#18

generalist · v1 · seit 2. April 2026 · 8× · zuletzt 30. Juni 2026

17

Momentum

GEN-1 is a fully from-scratch trained, multimodal embodied foundation model by Generalist AI designed for physical robot control. It was trained on over 500,000 hours of real-world physical interaction data collected primarily via low-cost wearable devices on humans — with no teleoperation or simulation data. According to the company, the model achieves 99% success rates on several manipulation tasks (up from 64% with predecessor GEN-0) and completes tasks approximately three times faster than prior state-of-the-art. GEN-1 is currently available to early-access partners; no public pricing or general API access has been announced.

Preuves (8)