--- title: "Video Tokenization | World Models Glossary | world-models.io" canonical: "https://world-models.io/en/glossary/video-tokenization/" description: "The process of encoding raw video frames into a sequence of discrete or continuous tokens that a transformer can model. Video tokenizers (such as those." author: "Bernard Grenat" generated: "2026-10-30" --- New: the Timeline is live. Track world model releases, papers, and benchmark updates in real time. The Knowledge Hub for AI World Models # Video Tokenization The process of encoding raw video frames into a sequence of discrete or continuous tokens that a transformer can model. Video tokenizers (such as those released with NVIDIA Cosmos) are critical building blocks of foundation world models. robotics model-based-rl simulation embodied-ai ## What is a world model? A world model is an AI system that learns a predictive representation of how an environment changes over time. It can use observations, actions, or language to estimate future states and possible outcomes before an agent acts. World models support planning in model-based reinforcement learning, robotics, embodied AI, autonomous driving, and interactive simulation. This site catalogs these systems with model records, research papers, benchmark context, and source links. Start with the [world models database](https://world-models.io/en/world-models-database/), compare systems on the [leaderboard](https://world-models.io/en/leaderboard/), or read the [Performance Index methodology](https://world-models.io/en/methodology/performance-index/). ## What Is Video Tokenization? Editorial definition preserved directly in static HTML. The process of encoding raw video frames into a sequence of discrete or continuous tokens that a transformer can model. Video tokenizers (such as those released with NVIDIA Cosmos) are critical building blocks of foundation world models. Video Tokenization is a glossary concept in the techniques layer of the world models knowledge base. ## Term Snapshot Static glossary definition snapshot for crawlers and no-JS readers. | Attribute | Value | | --- | --- | | Term | Video Tokenization | | Category | Techniques | | Definition | The process of encoding raw video frames into a sequence of discrete or continuous tokens that a transformer can model. Video tokenizers (such as those released with NVIDIA Cosmos) are critical building blocks of foundation world models. | | Related Models | 2 | | Related Research | 1 | ## Related Models | Model | Lab | Category | | --- | --- | --- | | [NVIDIA Cosmos](https://world-models.io/en/models/nvidia-cosmos/) | NVIDIA | Foundation World Model | | [IRIS](https://world-models.io/en/models/iris/) | Microsoft Research | Model-Based RL | ## Related Research | Topic | Summary | | --- | --- | | [Foundation World Models](https://world-models.io/en/research/foundation-world-models/) | How foundation world models such as Cosmos and Genie 2 bring large-scale learned simulation to robotics, autonomous driving, and physical AI. | ## Related Terms | Term | Category | Definition | | --- | --- | --- | | [Latent Space](https://world-models.io/en/glossary/latent-space/) | Architecture | A compressed, abstract representation of data learned by a neural network. In world models, the latent space encodes environment states in a compact form that captures essential dynamics while discarding irrelevant details. Models like DreamerV3 and PlaNet operate entirely in latent space for efficient planning. | | [Foundation Model](https://world-models.io/en/glossary/foundation-model/) | Paradigms | A large-scale model trained on broad data that can be adapted to many downstream tasks. Foundation world models (like NVIDIA Cosmos and Genie 2) learn general-purpose representations of world dynamics, analogous to how GPT models serve as foundations for language tasks. | ## Quick Answer Short extractable summary preserved directly in static HTML. - Video Tokenization is a glossary concept used across world-models.io to clarify language, methods, and architectural ideas in the field. - Use this page to get the definition quickly, then continue into related models, research topics, and adjacent terms for context. ## 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 glossary page publishes stable definitions linked to related models, research topics, and primary-source context. 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: 2026-09-27. 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.](https://world-models.io/en/editorial-policy/) Primary sources only Last reviewed date visible Methodology documented Source links included Correction channel published Review date visible ## Reference Sources Primary sources related to this term, surfaced directly in static HTML. - [NVIDIA Cosmos paper](https://arxiv.org/abs/2501.03575) NVIDIA primary research source - [NVIDIA Cosmos official page](https://www.nvidia.com/en-us/ai/cosmos/) NVIDIA official page - [IRIS paper](https://arxiv.org/abs/2209.00588) Microsoft Research primary research source - [IRIS cited source](https://arxiv.org/abs/1803.10122) Micheli et al., 2023. Transformers are Sample-Efficient World Models. ICLR 2023.