RoboWits - Seed Tasks: leaderboard
RoboWits (Lin et al., 2026) is a simulated bimanual manipulation benchmark of creative problem solving: seed tasks need geometry, material or assembly reasoning, such as using a rod to reach an object or displacing water to raise its level. Success rate in percent over 10 seed tasks, 50 trials each with randomly placed objects, for manipulation policies trained or fine-tuned on 50 human demonstrations of each task; higher is better.
Metric: Task success rate (%; mean over the 10 evaluated seed tasks of 50 randomly placed trials each; bimanual simulation; policy trained or fine-tuned on the 50 human teleoperation demonstrations of each of the 10 seed tasks). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | π0.5 | 12.6 | |
| 2 | π0 | 12.2 | |
| 3 | ACT | 11 |
Interactive version: theaggregate.ai/benchmark?slug=robowits-seed-tasks · How It Works · Data refreshed daily, snapshot 2026-11-12.