Skip to content
RobotWorld
← Back

Riemann-1.0 trains a World Action Model on 200k hours of human video

Loading video
Loading video

Riemann Dynamics introduces Riemann-1.0, a World Action Model trained on more than 200,000 hours of egocentric human video combined with gripper data and robot trajectories, targeting general-purpose robot manipulation, world simulation, and action generation.

World Action Model世界动作模型具身智能Embodied AI人类视频Human Video机器人操作ManipulationRiemann DynamicsRiemann-1.0Twitter
Category: research
Author: @techniahqrobot
Date: 2026-08-29T00:00:00
Duration: 22.2s
Source: https://x.com/techniahqrobot/status/2093647299196973300
Reference: https://arxiv.org/abs/2608.27033

Related Videos

τ0-VLA: Robot VLA that searches over subtask sequences at test time with a world model
3:33

τ0-VLA: Robot VLA that searches over subtask sequences at test time with a world model

@jiqizhixin · 43 views · 2026-08-07
VLA世界模型World Model
Nori SDK is an open-source Python and TypeScript robotics development kit
1:06

Nori SDK is an open-source Python and TypeScript robotics development kit

@AntonioSitongLi · 45 views · 2026-08-28
机器人 SDKRobot SDK开源
Thea: Towards the Harness of Embodied Agents
1:06

Thea: Towards the Harness of Embodied Agents

@walterzhu8 · 14 views · 2026-08-13
具身智能Embodied AIAgent
FACT: Learning from failure in robot learning
2:17

FACT: Learning from failure in robot learning

@QuanquanPeng03 · 11 views · 2026-08-12
具身智能Embodied AI机器人学习
Robot prototype with full-body flexible electronic skin
0:50

Robot prototype with full-body flexible electronic skin

@warmcore7278 · 13 views · 2026-08-11
电子皮肤Electronic Skin触觉感知
BeingBeyond: Beijing embodied-AI startup building a pure robot brain
2:34

BeingBeyond: Beijing embodied-AI startup building a pure robot brain

@ErenChenAI · 11 views · 2026-08-08
BeingBeyond具身智能Embodied AI