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

Genie vs Genie 2

Genie pioneered unsupervised interactive environment generation from video. Genie 2 massively scales this approach to generate persistent, interactive 3D worlds from single images.

robotics model-based-rl simulation embodied-ai

Comparison Overview

Main comparison summary preserved directly in static HTML.

Genie pioneered unsupervised interactive environment generation from video. Genie 2 massively scales this approach to generate persistent, interactive 3D worlds from single images.

Verdict

Primary editorial conclusion preserved for non-JS crawlers and readers.

Genie 2 is the natural evolution of Genie, moving from 2D to 3D, from basic movement to rich physics, from low to high resolution. Genie v1 was the proof-of-concept that interactive environments could be learned from unlabeled video. Genie 2 is the production-grade successor that could transform AI agent training.

Key Differences

Extractable difference list generated from the comparison table.

  • Environment Type: Genie (v1) - 2D platformer-style; Genie 2 - Full 3D environments.
  • Training Data: Genie (v1) - 200K hours internet video; Genie 2 - Curated 3D video data.
  • Input: Genie (v1) - Single image; Genie 2 - Single image.
  • Action Discovery: Genie (v1) - Latent actions from video; Genie 2 - Action-conditioned generation.
  • Physics: Genie (v1) - Basic (2D movement); Genie 2 - Object permanence, gravity, collisions.

When To Use Each

Static decision guidance for no-JS readers.

Choose Genie (v1) when...

Choose Genie (v1) when its capabilities best match your research or deployment requirements.

Choose Genie 2 when...

Choose Genie 2 when its capabilities best match your research or deployment requirements.

Comparison Table

Genie 2 is the natural evolution of Genie, moving from 2D to 3D, from basic movement to rich physics, from low to high resolution. Genie v1 was the proof-of-concept that interactive environments could be learned from unlabeled video. Genie 2 is the production-grade successor that could transform AI agent training.

DimensionGenie (v1)Genie 2
Environment Type2D platformer-styleFull 3D environments
Training Data200K hours internet videoCurated 3D video data
InputSingle imageSingle image
Action DiscoveryLatent actions from videoAction-conditioned generation
PhysicsBasic (2D movement)Object permanence, gravity, collisions
ResolutionLow resolutionHigh resolution 3D
Year2024 (early)2024 (late)

Performance Index Snapshot

High-level scoring context for the models referenced in this comparison.

ModelCategoryIndex v1.1Confidence
GenieGenerative World Model57/100medium
Genie 2Generative World Model79/100medium
OASISGenerative World Model66/100medium
SoraGenerative World Model63/100medium

Frequently Asked Questions

FAQ answers rendered directly into static HTML for extractable responses.

Do I need to understand Genie v1 before Genie 2?

Understanding Genie v1's latent action discovery and unsupervised learning approach provides useful context, but Genie 2 is architecturally distinct enough to study independently.

Quick Answer

Short extractable summary preserved directly in static HTML.

  • Genie (v1) vs Genie 2: this page compares where each system is stronger instead of forcing a universal winner.
  • Use the verdict for the short answer, then validate the trade-offs in the table, evidence sources, and benchmark context.
  • Related models and source links help connect this comparison to the broader world models landscape.

Editorial Trust Signals

Editorial provenance and refresh policy preserved directly in static HTML.

Published by world-models.io editorial board.

Lead editor Bernard Grenat.

This comparison page publishes a direct answer, explicit trade-offs, and source-backed evidence that can be validated against primary materials.

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: 2026-06-21.

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 papers and official sources for the models discussed on this comparison page.