---
title: "X-World 1.6: Building Controllable Virtual Cities with World Models"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2611.00596v1"
permalink: "https://feed.astralor.com/items/itm_5f8c5971f2f4afcff310ac54"
published: "2026-10-31T03:45:22.000Z"
score: 55
channel: "tech"
tags: ["安全","多模态","X-World","Building Controllable Virtual","Cities","World Models Deploying","Deploying"]
---

# X-World 1.6: Building Controllable Virtual Cities with World Models

## 速览

Deploying generative world models for end-to-end closed-loop evaluation and policy learning requires precise scenario control, consistent observations, and sustained generation quality during extended interactions. This…

## 关键点

- Deploying generative world models for end-to-end closed-loop evaluation and policy learning requires precise scenario control, consistent observations, and sustained generation quality during extended interactions.
- This report presents \emph{X-World~1.6}, advancing X-World~1.0 toward a more reliable and versatile driving simulator.
- The upgraded base model offers finer-grained control over traffic semantics, including traffic lights and road markings, together with improved spatial consistency across cameras.

## 全文

Deploying generative world models for end-to-end closed-loop evaluation and policy learning requires precise scenario control, consistent observations, and sustained generation quality during extended interactions. This report presents \emph{X-World~1.6}, advancing X-World~1.0 toward a more reliable and versatile driving simulator. The upgraded base model offers finer-grained control over traffic semantics, including traffic lights and road markings, together with improved spatial consistency across cameras. Virtual City capabilities broaden scenario construction through a family of models conditioned on diverse 2D and 3D references, including support for generating multi-camera future videos from a single image. An enhanced post-training framework enables efficient causal, few-step generation with stronger temporal coherence, better quality retention over longer rollouts, and improved alignment with human preferences. We also introduce a dedicated evaluation suite to support systematic assessment and guide model iteration. Together, these upgrades expand the scenarios X-World can simulate and improve the efficiency and reliability of its generated observations, supporting broader use in autonomous-driving development.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_5f8c5971f2f4afcff310ac54