ABot-Earth 0.7 – The World’s First 3D-Native Urban World Model Launches by Gaode

ABot-Earth 0.7 – The World’s First 3D-Native Urban World Model Launches by Gaode

2026-09-11 16:44

Introduction: Over one trillion daily calls to BeiDou, generating city-scale 3D models in 10 minutes. What is ABot-Earth 0.7? ABot-Earth 0.7 is a global first 3D-native urban world model launched by Amap.

What is ABot-Earth 0.7?

ABot-Earth 0.7 is the world’s first 3D-native urban world model introduced by Amap. Built on a 3D-native architectural framework, the model is trained using spatiotemporal data. Inputting a single satellite image or textual description generates a kilometer-scale 3D city within 10 minutes on consumer-grade GPUs. The model supports continuous, full-scale generation from planetary to street-level views, covering over 196 countries and regions, enabling users to freely explore and interact in real time. ABot-Earth 0.7 has been integrated into Amap’s Flight Street View 2.0, serving as the gateway to Amap’s "Spatial Intelligence" and evolving toward dynamic perception and spatiotemporal forecasting.

ABot-Earth 0.7

Key Features of ABot-Earth 0.7

  • Image-to/Text-to 3D City Generation: Generate kilometer-scale 3D city scenes in 10 minutes from a single satellite image or text prompt.
  • Full-Scale Continuous Generation: A single model seamlessly generates content from planetary scale down to street-level landmarks; spatial structure and visual fidelity remain consistent during zoom-in transitions.
  • Free Exploration & Real-Time Interaction: Autonomously completes continuous 3D space, enabling seamless entry, free roaming, and real-time interaction—delivering an immersive "presence effect."
  • Plug-and-Play Integration: Generated results integrate seamlessly with UE/Unity, facilitating developer adoption.
  • Flight Street View 2.0 Deployment: Users can enter complex buildings and large scenic areas, navigating 3D environments via joystick control.

Technical Principles of ABot-Earth 0.7

  • 3D-Native Architecture: Trained on spatiotemporal data that encodes both spatial and temporal information, enabling the model to develop native understanding of 3D space.
  • End-to-End 3DGS Generation: Outputs complete 3D Gaussian Splatting (3DGS) scenes in a single step, eliminating the need for multi-stage reconstruction.
  • Two-Stage Generation Pipeline: First generates sparse occupancy layout, then refines geometry and attributes (active voxels), progressively enhancing detail conditioned on satellite imagery.
  • Efficiency Breakthrough: Generates kilometer-scale scenes in 10 minutes on consumer-grade GPUs—efficiency improved by approximately 1,000x compared to traditional workflows.

How to Use ABot-Earth 0.7

  • Official Website Experience: Visit the official site at https://abot-earth.amap.com/, register and log in to experience it online.
  • Planet Exploration: Open the planet page on the website to freely browse Earth, Moon, Mars, and generated 3D urban spaces.
  • Image-to/Text-to 3D Generation: Upload a satellite image or input a textual description to generate a kilometer-scale 3D city scene.
  • Plot Square: Browse and roam 3D plots shared by others, experiencing the quality of generation and interactive immersion firsthand.
  • Use within Amap Maps: Access complex buildings and major scenic areas through Amap App’s “Flight Street View 2.0,” controlling navigation via joystick in 3D environments.
  • Engine Integration (for Developers): Directly import generated 3DGS outputs into Unity or Unreal Engine, add interaction logic, and deploy for use in gaming, film production, or digital twin applications.

Core Advantages of ABot-Earth 0.7

  • World’s First 3D-Native Architecture: Establishes native understanding of 3D space at the foundational level—a paradigm-shifting innovation in digital earth technology.
  • Exceptional Generation Efficiency: Generates kilometer-scale 3D cities in 10 minutes on consumer-grade GPUs, achieving ~1,000x efficiency improvement over legacy methods.
  • Full-Scale Continuous Generation: One unified model spans planetary → city → street-level landmarks; structural and visual consistency maintained when zooming from macro to micro scale.
  • Unmatched Global Coverage: Covers over 196 countries and regions—the broadest coverage currently available in digital earth platforms.
  • Freedom in Interaction Experience: Autonomously completes continuous 3D space, enabling free roaming and real-time interaction—free from fixed routes typical of traditional street view systems.
  • Deep Data Moats: Leverages ten years of accumulated spatiotemporal data from Amap’s daily near-trillion calls to BeiDou positioning, creating irreplicable real-world training samples.
  • Ecosystem Plug-and-Play: Seamlessly integrates with UE/Unity and is deployed in Amap Flight Street View 2.0, offering a complete path from user experience to commercial application.

Competitive Comparison: ABot-Earth 0.7 vs. Alternatives

Comparison Dimension ABot-Earth 0.7 (Amap) Google Earth / 3D Tiles (Google)
Technology Path 3D-Native Generation: Model understands 3D space, end-to-end generation of 3DGS scenes Imagery Reconstruction: Stitching satellite images, aerial photography, and point cloud data
Content Source AI-generated + autonomous completion, unconstrained by acquisition materials Dependent on real-world footage and acquisition paths; unvisited areas cannot be viewed
Generation Speed Generates kilometer-scale 3D cities in 10 minutes—~1,000x faster than traditional methods Requires professional acquisition and months-long reconstruction pipelines; no instant generation capability
Coverage Scope Over 196 countries and regions (scalable via AI generation) High-definition coverage of major global cities; low accuracy in remote areas
Interaction Method Free entry, continuous roaming, real-time interaction—with immersive "presence" Browsing-centric; view angles and paths constrained by source material
Openness Level Generated outputs integrate with UE/Unity, open to developer ecosystem 3D Tiles offers API, but generation capabilities are not open to third parties
Strategic Positioning Gateway to Spatial Intelligence, evolving toward dynamic perception and spatiotemporal inference Geospatial browsing and map service platform

Application Scenarios of ABot-Earth 0.7

  • Smart Mobility Navigation: Enable 3D real-world navigation and pre-visit exploration in Amap Maps or in-car systems—step into scenic spots, commercial zones, and complex buildings before departure.
  • Autonomous Driving Simulation Training: Rapidly generate high-fidelity, kilometer-scale urban environments for cost-effective, customizable simulation testing of autonomous vehicles.
  • Game & Film Production: Generate 3D city assets in minutes and integrate into UE/Unity, drastically reducing costs for open-world games and virtual set design in film production.
  • Digital Twins & Smart Cities: Build real-time interactive 3D city foundations for urban planning, traffic management, and emergency simulation exercises.
  • Cultural Tourism & Real Estate Showcase: Enable 3D online exploration of scenic areas and museums, and serve as spatial perception training data for embodied robots and smart glasses.

Source: AI Tools Hub

#AI Tools

Disclaimer: Contains third-party opinions, does not constitute financial advice

Share To
X
Telegram
WeChat
QQ
Link
Recommended Reading

Hy4 Preview Lightweight Version – Tencent Hunyuan's Open-Source Hy4 Quantized Compression Model

10 days ago
Hy4 Preview Lightweight Version – Tencent Hunyuan's Open-Source Hy4 Quantized Compression Model

Claude Fable 5.1 – Anthropic's Latest Flagship Large Model

10 days ago
Claude Fable 5.1 – Anthropic's Latest Flagship Large Model

Open-source Retool is here, drag-and-drop backend with 60+ components

24 days ago
Open-source Retool is here, drag-and-drop backend with 60+ components

Claude Opus 5 is Here: Half the Price Approaching Fable 5

07-25
Claude Opus 5 is Here: Half the Price Approaching Fable 5

Tencent Open-Sources 260 Real-World Tasks to Test AI Workforce Capabilities

07-25
Tencent Open-Sources 260 Real-World Tasks to Test AI Workforce Capabilities

Meituan Open-Sources AI Exam Platform: 813 Minecraft Tasks

07-25
Meituan Open-Sources AI Exam Platform: 813 Minecraft Tasks

Microsoft's Image Model Integrated into PPT, GPU Costs Reduced by 84%

07-24
Microsoft's Image Model Integrated into PPT, GPU Costs Reduced by 84%