88Z AI, 9 June — Amap has released ABot-Earth 0.5, the first 3D-native urban world model. It supports multimodal inputs and generates high-fidelity scenes quickly on consumer-grade GPUs, covering numerous countries and regions worldwide.
At a glance
- ABot-Earth 0.5 is the first 3D-native urban world model supporting multimodal inputs.
- A single consumer-grade GPU can generate a 1 sq km high-fidelity 3D scene in 10 minutes.

The update
Amap has officially launched ABot-Earth 0.5, the industry's first 3D-native urban world model. This model features robust multimodal processing capabilities, accepting various inputs such as images and text. Technically, it is built on 3D Gaussian Splatting representation and employs a native LoD decoder along with a sliding window seamless inference strategy, significantly enhancing generation efficiency and quality. Thanks to these innovations, users can generate high-fidelity 3D urban scenes covering one square kilometer in just 10 minutes using a single consumer-grade GPU. Currently, the model's services extensively cover over 190 countries and regions worldwide, involving more than 300 cities, providing efficient technical support for digital city construction.
Sources & editorial note
This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.
- 高德技术Source published
Source dates above refer to the original announcements; the date under the headline is this article’s publication on 88Z AI.