New: the Timeline is live. Track world model releases, papers, and benchmark updates in real time.
world-models.io
The Knowledge Hub for AI World Models

Astra

Astra is a native interactive world model evaluated by WBench.

robotics model-based-rl simulation embodied-ai

Key Attributes

Static HTML snapshot of the model record for crawlers and no-JS readers.

AttributeValue
ModelAstra
Lab / OrganizationTsinghua
CategoryGenerative World Model
SubtypeAction-Conditioned Interactive World Model
World Model TypeAction-Conditioned Interactive World Model
Primary DomainInteractive Gaming
ArchitectureAction-conditioned autoregressive video world model
ModalityImage + Discrete Actions -> Interactive Video
Training MethodAutoregressive interactive world modeling
Statusemerging
Year2026

About Astra

Main editorial body preserved directly in static HTML.

Astra generates an evolving environment conditioned on action input. It is included as a native interactive system rather than a text-to-video model wrapped in WBench's chained-context protocol.

Astra is an action-conditioned interactive world model developed by Tsinghua University in 2026 for interactive gaming.

Editorial Snapshot

Short extractable facts for answer engines and no-JS readers.

SignalValue
DefinitionAstra is an action-conditioned interactive world model developed by Tsinghua University in 2026 for interactive gaming.
Short DescriptionAstra is a native action-conditioned interactive world model evaluated by WBench.
Benchmark Rows1
FAQ Entries1
Related Models2
Related Guides0
Related Research Topics0
Last Updated2026-08-31

Notable Features

Key capabilities associated with this model.

  • Discrete action control
  • Academic research
  • WBench evaluation

Use Cases

Representative applications attached to this model record.

Interactive environment simulationWorld-model evaluationInteractive Gaming

Strengths and Limitations

Balanced assessment surfaced in static HTML.

Strengths

  • Discrete action control
  • Academic research

Limitations

  • Early-stage performance
  • Primary project identity requires continued monitoring

Benchmarks

Published benchmark evidence attached to this model record.

BenchmarkMetricResultSource
WBenchFive-dimension average 63.7 /10063.7Source

References and Citations

Primary references preserved in static HTML for citation extraction.

ReferenceLink
Tsinghua University. Astra. Primary project source.Open source
Ying et al., 2026. WBench: A Comprehensive Multi-turn Benchmark for Interactive Video World Model Evaluation.Open source

Related Models

Nearby models linked from the current editorial record.

ModelCategoryWorld Model TypeIndex v1.1
MatrixGame3Generative World ModelAction-Conditioned Interactive Game World ModelN/A
HY-GameCraftGenerative World ModelAction-Conditioned Game World ModelN/A

Frequently Asked Questions

FAQ answers rendered directly into static HTML for extractable responses.

Why is Astra included?

Astra exposes native action conditioning and generates an evolving interactive environment; WBench evaluates that interface directly.

Quick Answer

Short extractable summary preserved directly in static HTML.

  • Astra is an action-conditioned interactive world model developed by Tsinghua University in 2026 for interactive gaming.
  • Use this page when you need a fast read on how Astra fits into the generative world model landscape, then validate the details in the benchmarks, citations, and related pages.
  • A key strength surfaced in the editorial record is discrete action control.

Editorial Trust Signals

Editorial provenance and refresh policy preserved directly in static HTML.

Published by world-models.io editorial board.

Lead editor Tyler D. - Technical editor, methodology and benchmark analysis.

This model page synthesizes primary papers, official model pages, benchmark evidence, and related world-models.io context into a reference resource.

Each editorial page is assembled from primary sources, normalized into extractable summaries, checked for factual drift, and reviewed before publication or major refreshes. Last reviewed: 2026-08-31.

Pages are refreshed when a new paper, benchmark, release, architecture update, or stronger primary source materially changes the answer a reader or AI system should retrieve.

Each page links back to relevant primary sources and keeps a stable canonical URL so readers can verify claims, trace context, and reference the most up-to-date version. See the editorial policy.

Primary sources onlyLast reviewed date visibleMethodology documentedSource links included

External Sources

Primary model and lab sources embedded in static HTML.

References

  1. [1] Tsinghua University. Astra. Primary project source.
  2. [2] Ying et al., 2026. WBench: A Comprehensive Multi-turn Benchmark for Interactive Video World Model Evaluation.