Model reference · open weights
WorldGuide-Ckpt is an open-weight video model from MBZUAI.
For self-hosting, compare this build's requirements below with the GPU you plan to rent.
WorldGuide-Ckpt is an image-to-video model developed by MBZUAI. It functions as a goal-directed video world model designed for procedural task execution. The model supports the English language.
Summary of the MBZUAI/WorldGuide-Ckpt model card, 2026-10-09
What it is
| Released by | MBZUAI |
|---|---|
| Released | 2026-10-02 |
From the model card
Ankan Deria, Komal Kumar, Hisham Cholakkal, Fahad Shahbaz Khan, Salman Khan
Mohamed bin Zayed University of Artificial Intelligence
Official checkpoints for WorldGuide. Method details, results and training code are on the project page and on GitHub.
WorldGuide-Ckpt/
├── transformer/ # WorldGuide Video DiT → --action_ckpt
└── text_encoder/llm/ # WorldGuide ContextPlanner → --planner_model_path
git clone https://github.com/mbzuai-oryx/WorldGuide.git
cd WorldGuide
pip install -r requirements.txt
# Download this repository and all other required components into ./ckpts
python download_models.py --hf_token
# Run closed-loop inference
bash scripts/inference/run_v5_qwen_closed_loop_memory.sh
@misc{worldguide,
title = {WorldGuide: Goal-Directed Video World Model for Procedural Task Execution},
author = {Deria, Ankan and Kumar, Komal and Cholakkal, Hisham and Khan, Fahad Shahbaz and Khan, Salman},
year = {2026}
}
Quoted from the model card on Hugging Face. The full card is behind the Hugging Face link above.
How it works