ReadingResearchRadarInvestment framework
Sign in / Sign up
← Back to reading
TopicDisconfirmAI EditorialUpdated 2026-11-22 01:53

Duo v1 Reciprocal World Model

3 steps1 pieces3 Sources

Ant Group's Technology Research Institute and Hangzhou-based AI company Westlake HeartChen have released and open-sourced Duo v1, a Reciprocal World Model (RWM) that integrates intelligence and physical reality into a single framework. Unlike traditional world models that simulate environments for external agents to control via keyboard or text, Duo v1 enables virtual agents to continuously perceive real-time audio and video inputs while autonomously deciding when to speak, listen, or interrupt. The model utilizes two coupled modules: Intent, which handles cognitive reasoning and interaction timing decisions in ~0.7-second chunks, and Realization, which generates synchronized audiovisual responses. In human evaluations across 30 real-world interaction scenarios, Duo v1 achieved a 93.3% pass rate on 180 checks, with perfect scores in language adherence, speech intelligibility, and visual identity stability. However, content responsiveness remains a challenge at 73.3%. The technical report, training code, and model weights are now publicly available.

How it got here

  1. 2026-11-11

    2026-11-11: The research team submitted the paper 'Duo: A Reciprocal World Model for Full-Duplex Interaction' to arXiv, introducing the RWM concept.

    arxiv.org ↗
  2. 2026-11-13

    2026-11-13: Ant Group and Westlake HeartChen officially released Duo v1 on GitHub and Hugging Face, providing model weights, inference code, and streaming pipelines.

    github.com ↗
  3. 2026-11-17

    2026-11-17: Media outlets including Zhidx.com reported on the public release and open-source availability of the Duo v1 model.

    zhidx.com ↗

What is still unsettled

Sign in / Sign up
ReadingResearchRadarInvestment framework
  • How will future versions address the current limitation in content responsiveness (73.3%)?
  • What specific improvements are planned for Duo v2 regarding direct visual interaction conditioning?

Sources

  • arxiv.orgarXiv Paper↗
  • github.comGitHub Repository↗
  • zhidx.comZhidx.com Report↗

Coverage in this topic

  • Ant Group and Westlake HeartChen Open-Source Duo v1: A Reciprocal World Model Where Virtual Agents Autonomously Decide When to Speak or Listen2026-11-19News

    Ant Group's Technology Research Institute and Hangzhou-based AI company Westlake HeartChen have released and open-sourced Duo v1, a Reciprocal World Model that enables virtual agents to autonomously manage interaction timing and generate real-time audiovisual responses.