AI2's MolmoAct 2 Tops Human-Preference Robot Rankings; 55K-Vote Dataset Open-Sourced
allen_ai · x · 2026-10-14
AI2 announced its robot policy model MolmoAct 2 ranked #1 in Datapoint AI's human-preference rankings for real-world manipulation tasks like candy sorting and cup stacking.
The eval is based on the open-source robot-policy-human-preferences-55k dataset (CC-BY-4.0, on Hugging Face):
- 54,940 validated pairwise votes, exactly 10 per pair across 5,494 real robot run pairs
- Same bimanual robot and matched starting layout per task
- Covers 5 VLA models including MolmoAct 2, OpenVLA-OFT, and π0.5
- Overall and category-wise rankings are public, with dataset and eval recordings fully open
More from Embodied
- DeLorean joins Airspeeder International Series, first American team debuting in November — evilrabbit_ · 2026-10-14
- Gemini 4 Argon taught Apptronik's Apollo humanoid cornhole from video, sinking 7 straight throws — GoogleDeepMind · 2026-10-14
- Box2D dev unveils contact recycling, making contact generation almost free in physics sims — EricLengyel · 2026-10-14
- Startup Squad Turns Idea Into Robot-Drawn Brooklyn Bridge Sketch in 30 Minutes — davidfromkansas · 2026-10-14
- Stream Ring, an AI wearable three years in the making, starts shipping next month — pzakin · 2026-10-14
- Dyna-2 Robots Predict the Future: Million-Hour World Models and the Bitter Lesson of Robotics — JasonMa2020 · 2026-10-14