Tsinghua-linked VPP2 world action model tops RoboDojo simulation leaderboard
Overview
Star Motion Era (星动纪元) says its self-developed world action model VPP2 ranked first on the RoboDojo simulation leaderboard, with a 32.26% average success rate and 39.26 average score, according to a QbitAI (量子位) report.
The report attributes the gains to staged training that separates video prediction from action learning.
The same report says VPP2 reached a 58.5% zero-shot success rate on a real ALOHA dual-arm robot, versus 40% for π0.5. It also states that the code is open source on GitHub, allowing others to try the model and reproduce the paper's results. These performance figures come from the company and the article; this overview does not independently verify them.
Written by AI from the articles below · updated Oct 9, 1:06 AM ET
Check the sources:
Article timeline
Follow the coverage from different perspectives. Times are ET.
- QbitAITsinghua-linked VPP2 world action model tops RoboDojo simulation leaderboard
AIStar Motion Era's VPP2, a world action model, ranked first on the RoboDojo simulation leaderboard with a 32.26% average success rate and 39.26 average score. The article attributes gains to staged training that separates video prediction from action learning, and reports a 58.5% zero-shot success rate on a real ALOHA dual-arm robot versus 40% for π0.5. The code is open source on GitHub.
Heat trend
Not enough continuous observations to show a trend yet.