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

IRIS

IRIS is an autoregressive world model that bridges language modeling and reinforcement learning, using discrete tokens to represent environment dynamics.

robotics model-based-rl simulation embodied-ai

Key Attributes

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

AttributeValue
ModelIRIS
Lab / OrganizationMSR
CategoryModel-Based RL
SubtypeAutoregressive World Model
World Model TypeDiscrete token-based dynamics model
Primary DomainAtari games
ArchitectureVQ-VAE tokenizer + autoregressive Transformer dynamics model
ModalityVisual
Training MethodDiscrete tokenization + autoregressive next-token prediction over (observation, action, reward) sequences
Statusactive
Year2023
Performance Index65/100 (medium confidence, v1.1)

About IRIS

Main editorial body preserved directly in static HTML.

IRIS (Imagination with auto-Regression over an Inner Speech) treats world modeling as a sequence modeling problem. It tokenizes visual observations into discrete tokens using a VQ-VAE, then models environment dynamics autoregressively using a Transformer. This bridges the gap between language modeling and world modeling, achieving strong Atari 100K results.

IRIS is a discrete token-based dynamics model developed by Microsoft Research in 2023 for atari games.

Editorial Snapshot

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

SignalValue
DefinitionIRIS is a discrete token-based dynamics model developed by Microsoft Research in 2023 for atari games.
Short DescriptionA world model agent that tokenizes observations into discrete tokens and models environment dynamics autoregressively, like a language model over game frames.
Benchmark Rows1
FAQ Entries1
Related Models2
Related Guides2
Related Research Topics3
Last Updated2026-02-25

Notable Features

Key capabilities associated with this model.

  • Treats world modeling as sequence modeling
  • Discrete token-based dynamics
  • Transformer architecture for dynamics
  • Bridges language modeling and world modeling

Use Cases

Representative applications attached to this model record.

Atari gamesSample-efficient RLWorld model research

Strengths and Limitations

Balanced assessment surfaced in static HTML.

Strengths

  • Simple and elegant architecture
  • Strong Atari 100K results
  • Connects to LLM methodology
  • Scalable with Transformer advances

Limitations

  • VQ-VAE tokenization may lose fine details
  • Autoregressive generation is slow
  • Limited to discrete token spaces

Benchmarks

Published benchmark evidence attached to this model record.

BenchmarkMetricResultSource
Atari 100KMean HNS 1.046 x humanStrong performanceSource

References and Citations

Primary references preserved in static HTML for citation extraction.

ReferenceLink
Micheli et al., 2023. Transformers are Sample-Efficient World Models. ICLR 2023.Open source

Related Models

Nearby models linked from the current editorial record.

ModelCategoryWorld Model TypeIndex v1.1
DreamerV3Model-Based RLImagination-based dynamics model88/100
MuZeroModel-Based RLAbstract learned dynamics + MCTS78/100

Direct Comparisons

Side-by-side comparisons already connected to this model.

ComparisonMatchupSummary
World Models vs LLMsWorld Models vs Large Language ModelsWorld models and LLMs represent fundamentally different approaches to AI. World models learn causal dynamics of physical environments; LLMs learn statistical patterns over text. Both are essential for the future of AI.
DreamerV3 vs DIAMONDDreamerV3 vs DIAMONDDreamerV3 and DIAMOND are both model-based RL agents that train policies via imagination, but they use fundamentally different dynamics models: RSSM latent dynamics vs. pixel-space diffusion models.
IRIS vs DreamerV3IRIS vs DreamerV3IRIS and DreamerV3 are both leading model-based RL agents but use fundamentally different world model architectures: autoregressive token prediction vs. RSSM latent dynamics.
DreamerV3 vs IRISDreamerV3 vs IRISTwo model-based RL agents using fundamentally different world model architectures: DreamerV3's RSSM with actor-critic vs. IRIS's autoregressive Transformer with VQ-VAE tokens.
IRIS vs DIAMONDIRIS vs DIAMONDTwo approaches to learning game simulators: IRIS uses discrete tokenization with a GPT-like transformer, while DIAMOND leverages diffusion models for higher visual fidelity.

Guides Referencing This Model

Crawler-readable guide links tied to this model.

GuideSummary
Building World Models: A Practical GuideA practical guide to implementing world models: from choosing architectures and training setups to debugging dynamics learning and policy optimization.
Evaluating World Models: Benchmarks, Metrics and PitfallsA practical reference on how to benchmark world models: from sample efficiency on Atari 100K and DMControl to long-horizon prediction quality and downstream policy performance.

Research Topics Referencing This Model

Connected research areas surfaced directly in static HTML.

TopicSummary
Model-Based Reinforcement LearningWhat model-based reinforcement learning is, how world models enable imagination-based planning, and why Dreamer, MuZero, PlaNet, and TD-MPC2 matter.
World Models vs LLMsThe key differences between world models and LLMs across objective, architecture, planning, physical reasoning, and embodied AI use cases.
World Model EvaluationHow to evaluate world models across rollout quality, benchmark performance, planning utility, and downstream transfer instead of relying on visual plausibility alone.

Frequently Asked Questions

FAQ answers rendered directly into static HTML for extractable responses.

How does IRIS relate to language models?

IRIS treats world modeling as sequence prediction over discrete tokens, similar to how LLMs predict next tokens in text, but applied to game observations and actions.

Quick Answer

Short extractable summary preserved directly in static HTML.

  • IRIS is a discrete token-based dynamics model developed by Microsoft Research in 2023 for atari games.
  • Use this page when you need a fast read on how IRIS fits into the model-based rl landscape, then validate the details in the benchmarks, citations, and related pages.
  • A key strength surfaced in the editorial record is simple and elegant architecture.

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: 2026-02-25.

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] Micheli et al., 2023. Transformers are Sample-Efficient World Models. ICLR 2023.