1X introduces NEO Gamma with revised home design and manipulation AI

1X has introduced NEO Gamma, a revision of its home humanoid combining a softer domestic design with changes to whole-body control and visual manipulation. Announced on 21 February, Gamma follows the NEO Beta prototype and is intended to support internal home testing.

The controller runs at 100 Hz and uses reinforcement learning from human motion-capture data. Demonstrated behaviours include walking with natural arm swings, squatting to reach the floor and sitting in chairs while maintaining balance.

For manipulation, 1X has trained neural networks to predict teleoperated actions from raw sensor data. The company says its visual model can pick up varied objects in different settings, including environments absent from training.

Soft covers surround the tendon-drive actuation system, while a seamless knit suit and illuminated ear rings support a more approachable appearance and visual feedback. 1X also reports lower operating noise and improved hardware reliability. Gamma remains a development platform combining autonomy and teleoperation, with home testing intended to guide further iteration.