Skip to content
RobotWorld
← Back to Videos

Odyssey-3: A Foundation World Model for Physical Agents

Loading video
Loading video

Odyssey-3 is an autoregressive diffusion-transformer foundation world model that controls robot arms, humanoids, cars, and drones with task-specific experiential data.

World ModelsFoundation ModelsPhysical AIRobot LearningHumanoid robotsRobot ArmsAutonomous DrivingDrone NavigationEmbodied AIDiffusion TransformerOdysseyHumanoid
Category: research
Author: @odysseyml
Date: 2026-09-15T00:00:00
Duration: 96.833s
Source: https://x.com/odysseyml/status/2099900067356586276
Reference: https://odyssey.systems/introducing-odyssey-3

Related Videos

Being-H0.8: Latent Tactile World Action Model
2:32

Being-H0.8: Latent Tactile World Action Model

@beingbeyond_ · 105 views · 2026-07-28
Tactile SensingWorld ModelsEmbodied AI
LingBot-World 2.0 Opens Infinite Worlds
2:04

LingBot-World 2.0 Opens Infinite Worlds

@robbyant_brain · 124 views · 2026-07-08
World ModelsEmbodied AIInteractive Video
FACT: Learning from failure in robot learning
2:17

FACT: Learning from failure in robot learning

@QuanquanPeng03 · 130 views · 2026-08-12
Embodied AIRobot LearningFailure Learning
Robot prototype with full-body flexible electronic skin
0:50

Robot prototype with full-body flexible electronic skin

@warmcore7278 · 106 views · 2026-08-11
Electronic skinTactile PerceptionHumanoid robots
BeingBeyond: Beijing embodied-AI startup building a pure robot brain
2:34

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

@ErenChenAI · 91 views · 2026-08-08
BeingBeyondEmbodied AIFoundation Models
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 · 213 views · 2026-08-07
VLAWorld ModelsEmbodied AI