Research · 9 Oct 2026 · Company announcement

Robotera says its VPP2 model tops the RoboDojo manipulation benchmark

Purpose: sending the RankedRobot digest. Nothing else; never sold or shared for marketing.
What we keep: your email address, the date you signed up and the page you signed up on.
How long: until you unsubscribe, when the address is deleted. Before the first issue we email you once to confirm; addresses not confirmed within 30 days are deleted.
Who processes it: stored with Cloudflare, our host, which may process it outside your country. We'll name our email-sending provider in the privacy notice before the first issue.
Your choice: you don't have to subscribe; the only effect of saying no is that you won't get the digest. Unsubscribe any time.

The Beijing company reports a 32.26% average success rate and has released code and checkpoints; the result is self-reported.

Robotera says its Video Prediction Policy 2 (VPP2) ranked first on RoboDojo, a benchmark built by the University of Hong Kong's MMLab with nearly 20 academic institutions. The company reports a 32.26% average success rate and an average score of 39.26, and first place in the generalisation, precision and memory categories.

VPP2 is described as a world-action model: it predicts how actions will change a scene and turns instructions into movements, optionally with a vision-language model that breaks tasks into steps. Robotera also reports 58.5% average success across ten ALOHA task categories and 45.0% on LIBERO-Pro.

Robotera says it used no additional data or agent-based self-improvement. The code is open-sourced on GitHub. The release does not say how many methods were evaluated, and the results have not been independently verified.

Sources

Written by the RankedRobot news desk (AI-assisted, human-reviewed) from the sources above; checked 2026-10-12. Claims are attributed to whoever makes them. Spotted an error? corrections@rankedrobot.com