# LingBot-World

> LingBot-World is an open-source real-time interactive world model that generates infinite explorable 3D environments from a single image. Developed by Ant Group's Lingbo Technology (Robbyant), it rivals Google Genie 3 in quality but is fully open-source under Apache 2.0 license.

LingBot-World represents a paradigm shift from passive video generation to active world simulation. With 28 billion parameters and 16 FPS real-time generation, LingBot-World delivers sub-second latency and stable long-term memory for 10+ minutes of continuous exploration. The model supports multiple visual styles from photorealistic to anime, making it versatile for game development, embodied AI training, autonomous driving simulation, and creative content generation.

## Core Features

- [Long-term Memory](https://lingbot-world.top/#features): Maintains consistent environments for 10+ minutes without "ghost walls" problem
- [Style Generalization](https://lingbot-world.top/#features): Supports photorealistic, anime, cartoon, game, and fantasy visual styles
- [Action Agent](https://lingbot-world.top/#features): AI agent with autonomous navigation and WASD keyboard controls

## Technical Specifications

- [Architecture](https://lingbot-world.top/#technical): 28B parameters (14B inference), Diffusion Transformer + MoE
- [Performance](https://lingbot-world.top/#technical): 16 FPS, sub-second latency, 480P/720P output
- [Training Data](https://lingbot-world.top/#technical): Real videos + Game recordings + Synthetic scenes

## Model Versions

- [LingBot-World-Base (Cam)](https://lingbot-world.top/#models): Camera pose control, available now
- [LingBot-World-Base (Act)](https://lingbot-world.top/#models): Action control, coming soon
- [LingBot-World-Fast](https://lingbot-world.top/#models): Real-time streaming, coming soon

## Getting Started

- [Quick Start Guide](https://lingbot-world.top/#getting-started): Deploy LingBot-World in 4 simple steps
- [GitHub Repository](https://github.com/Robbyant/lingbot-world): Source code and documentation
- [Research Paper](https://github.com/Robbyant/lingbot-world/blob/main/LingBot_World_paper.pdf): arXiv:2601.20540

## Resources

- [Hugging Face](https://huggingface.co/collections/robbyant/lingbot-world): Model weights and demos
- [ModelScope](https://www.modelscope.cn/collections/Robbyant/LingBot-World): Chinese mirror for model downloads
- [FAQ](https://lingbot-world.top/#faq): Common questions about LingBot-World

## Optional

- [Privacy Policy](https://lingbot-world.top/#privacy-policy): Data collection and usage policies
- [Terms of Use](https://lingbot-world.top/#terms-of-use): Apache 2.0 license terms
- [Disclaimer](https://lingbot-world.top/#disclaimer): Community information site notice