# Alibaba — Qwen-RobotWorld: Boundless Worlds for Embodied Agents

- Company: Alibaba (alibaba.com)
- Announced: 2026-06-16T00:00:00+00:00
- Category: research-paper
- Subject: Qwen
- Models affected: Qwen-RobotWorld, Qwen2.5-VL
- Source: https://qwen.ai/blog?id=qwen-robotworld
- Record: https://forck.live/items/4613-qwen-robotworld-boundless-worlds-for-embodied-agents

Alibaba introduces Qwen-RobotWorld, a dual-stream diffusion world model that uses natural language as a universal action interface to unify 20+ robot embodiments and 500+ action categories, enabling cross-scenario physical generalization across manipulation, driving, and navigation. The model employs Qwen2.5-VL as an action encoder within an MMDiT architecture, and is trained on the Embodied World Knowledge (EWK) dataset of 8.6M video-text pairs. It also supports multi-view geometrically consistent generation and human-to-robot transfer via Scene2Robot.

## Evidence

Verbatim from https://qwen.ai/blog?id=qwen-robotworld:

> Qwen-RobotWorld bridges this gap by treating natural language as a universal action interface. ... Language unifies the action space : world knowledge and embodied knowledge reinforce each other within a single model, enabling cross-scenario, cross-task physical generalization .

---

Record: https://forck.live/items/4613-qwen-robotworld-boundless-worlds-for-embodied-agents
Catalogue: https://forck.live/llms.txt
Feed: https://forck.live/feed.md