Ai2 releases Molmo2-4B-tb, a vision-language planner for robot block stacking
DJiafei · x · 2026-11-06
Ai2 released Molmo2-4B-tb on Hugging Face: a vision-language planner for robot block stacking that reads a goal, looks at the camera feed, and picks the next action by pointing.
A demo shows VLAGent Goal Conditioning in action: given a 'stack W G C' prompt, Molmo2 updates its output at 2Hz, pointing to the target block and issuing low-level instructions (pick with coordinates) executed by a motion planner.
More from Embodied
- Developer Validates AME-2 on Humanoids: 70cm Foot-Only Climbing, 1.2m Gap Traversal in Sim — ChongZzZhang · 2026-11-06
- Wayve releases Learned Safety strategy after a decade of end-to-end AV research — alexgkendall · 2026-11-06
- Generalist AI demos GEN-1.5 with millimeter-scale O-ring dexterity tasks — ChongZzZhang · 2026-11-06
- Florida deploys robot bunnies to lure invasive pythons that wiped out 99% of raccoons — IsForAt · 2026-11-06
- Lamport Robotics' Pincer1: a Six-Legged Robot Wallet That Signs Only When You're in the Room — R4nsack · 2026-11-06
- UCLA's FlowPilot uses human preferences to cut sidewalk robot interventions by 40% — zhoubolei · 2026-11-06