What this lets people do
See how world models move from generated scenes into robots, sensors, and physical decision-making.
Embodied reasoning, spatial understanding, task planning, multimodal perception, and VLA-adjacent robot-control workflows.
Embodied foundation model
HY-Embodied-0.5 world model
See how world models move from generated scenes into robots, sensors, and physical decision-making.
Tencent Robotics X / HY Vision TeamOpen-source model family with enhanced 0.5-X releaseGitHub repository, arXiv paper, Hugging Face weights, and an April 2026 HY-Embodied-0.5-X model release.
Physical AIScene explainer
Three frames before the source list.
The page starts with the experience, then moves toward source-backed details.
01First impression
A visible world
HY-Embodied-0.5 is Tencent's embodied foundation-model family for spatial reasoning, embodied planning, and robot-facing perception, later extended with the HY-Embodied-0.5-X enhancement.
02Capability
Why it stands out
Adds a primary-source Tencent anchor in the site's physical-AI lane without pretending every embodied model is itself an interactive world simulator.
03Boundary
What not to overclaim
HY-Embodied-0.5 is an embodied foundation and planning model family, not a consumer world-creation product like Marble or HappyOyster.
Good reasons to open this page
- Readers tracking Tencent's physical-AI and robot-facing model lane beside HY-World 2.0.
- Comparisons that need to separate embodied reasoning from generated 3D world creation.
- Source-backed checks of open model availability, checkpoints, and the later 0.5-X enhancement.
Strengths
- Adds a primary-source Tencent anchor in the site's physical-AI lane without pretending every embodied model is itself an interactive world simulator.
- Public release surfaces are strong and reproducible: official GitHub code, arXiv paper, and Hugging Face checkpoints are all available.
- Useful for comparing world-model-adjacent robot cognition with Tencent's separate HY-World 2.0 generated-3D-world stack.
Limits and source boundary
- HY-Embodied-0.5 is an embodied foundation and planning model family, not a consumer world-creation product like Marble or HappyOyster.
- Its public claims center on embodied benchmarks and robot-facing reasoning, so it should not be framed as proof of a generally available world simulator.
Evidence and update history
Primary-source embodied-AI dossier with GitHub repository, arXiv paper, Hugging Face weights, and a dated 0.5-X release signal.
- 2026-04-09 · First tracked sourceHY-Embodied-0.5 entered the site as a embodied foundation model from Tencent Robotics X / HY Vision Team.
- 2026-04-24 · Latest dossier reviewThe page was reviewed for access status, source confidence, category boundary, and related comparison links.
- 2026-04-24 · Physical AITencent added the HY-Embodied-0.5-X release path on top of HY-Embodied-0.5, widening the robot-facing model surface.
Use it for, not for
Use it for
- HY-Embodied-0.5 is valuable because it gives Tencent coverage a second lane beyond generated 3D worlds: spatial reasoning and embodied planning.
- The page should help readers decide whether they are looking at a world generator, a robot-facing foundation model, or a benchmarked planning system.
- The 0.5-X update is treated as a release signal, not as blanket proof that every downstream robot workflow is production-ready.
Do not use it for
- Choosing a consumer AI environment creator, game-asset generator, or 360 skybox workflow.
- Claiming a complete robotics product or general simulator without matching the cited benchmark and release scope.
Quick workflow
- Start with the GitHub repository to identify supported model variants and evaluation setup.
- Use the Hugging Face page to verify the 0.5-X checkpoint surface.
- Use the HY-World 2.0 comparison path when the question shifts from embodied reasoning to generated 3D worlds.
Release signals
Only the selected updates that affect this profile.
The company profile stays stable. These short signals explain what changed and point back to sources.
Sources
FAQ
Dossier FAQ
Use these notes to keep model comments grounded in official sources and careful category boundaries.
What does World Models Watch count as a world model?
The site tracks systems that model environments, actions, spatial structure, or persistent simulated state. Pure text chatbots and ordinary video generators are only included when they provide a clear bridge toward interactive or physical world modeling.
Why do some AI video systems appear on a world-model site?
Video models are included only when they help explain the path from generated clips to controllable spaces, physics-aware prediction, or agent-ready simulation. The site keeps that distinction explicit so video generation is not overstated as a finished world simulator.
How does the site decide whether a release is reliable enough to list?
Primary sources carry the most weight: official product pages, research posts, papers, documentation, code repositories, and company announcements. Secondary media can be referenced, but it stays labeled as reported or adjacent unless independently confirmed.
What should readers post in comments?
Useful comments add source links, corrections, release-status notes, comparison questions, or concrete reader context. Comments are public immediately, so readers should avoid private information and unsupported promotional claims.
Discussion
Reader discussion
Add source-backed corrections, questions, or notes for this page.
Loading discussion...
Loading comments...