Model family

ZERITH

Participating organizations

Version scope and roles for these associations await verification.

View evidence

ZERITH Z-Lab’s video–action model uses adapted Wan2.2 features for robot manipulation. It combines multiview images, instructions and robot state to predict action chunks. Simulation policies and an H1 mid-training backbone are separate artifacts; their weights have noncommercial terms.

Overview

Model type
Action policy
Architecture
WAM
Action generation
Flow matching
Listed versions
ZERITH 2.5

Papers & reports (2)

ZERITH V2.5: From Predictive Power to Policy Learnability

Paper available

arXiv: 2611.23746
Authors & affiliations (21)

Author order: Original order

  1. Chenxi Han

    Not provided

  2. Ruby

    Not provided

  3. Epoch

    Not provided

  4. Orion

    Not provided

  5. Luke

    Not provided

  6. Bin

    Not provided

  7. Dusk

    Not provided

  8. Lyra

    Not provided

  9. Cal

    Not provided

First published
2026-11-18
Latest paper revision
v1 · 2026-11-18
Venue
Not provided

Covered model versions

  • zerith · ZERITH 2.5

    Exact ZERITH V2.5 preprint, published after the official technical report with its own author list

Paper ↗PDF ↗

ZERITH V2.5: From Predictive Power to Policy Learnability

Technical report

Authors & affiliations (1)

Author order: Team authorship

  1. ZERITH Z-Lab

    Not provided

First published
2026-11-15
Latest paper revision
v1.0
Venue
Not provided

Covered model versions

Paper ↗PDF ↗

Official resources

Submit a source or correction