LesenForschungRadarAnlageframework
Anmelden / Registrieren
← Zurück zur Lektüre
NewsThema · Duo v1 Reciprocal World ModelWesentlich2 Min. LesezeitDisconfirmAI Redaktion

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

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.

2026-11-18 23:00 GMT+8

According to a report by Zhidongxi on November 18, Ant Group's Technology Research Institute and Hangzhou-based large model company Westlake HeartChen released and open-sourced the Reciprocal World Model (RWM) Duo v1 on November 17. The model integrates an intelligent agent into the world through two tightly coupled modules: Intent and Realization. It allows virtual agents to continuously receive user audio and video inputs, autonomously decide when to speak or listen, and generate corresponding sounds and visuals in real time for interaction.

Traditional world models typically simulate only physical scenes such as lighting, materials, and dynamics, lacking an internally responsive agent. Duo v1 covers both physical reality and the intelligent agent, sharing the same interactive spatiotemporal space. Within each approximately 0.7-second chunk, the Intent module autonomously decides whether to speak or wait based on user input and interaction history, generating corresponding language content. The Realization module then combines the hidden state from Intent, user input, and its own generated historical content to produce matching audio and video.

In human evaluations across 30 sets of real interactions, Duo v1 passed 168 out of 180 checks, achieving an overall pass rate of 93.3%. Specifically, the pass rates for language adherence, speech intelligibility, and visual identity stability were all 100%, while content responsiveness was 73.3%. Content responsiveness and long-term consistency remain the main challenges for the current version. The technical report, training code, and model weights for Duo v1 are now publicly available for developers to download and build upon.

Anmelden / Registrieren
LesenForschungRadarAnlageframework

Quellen

  • zhidx.com↗
Thema

Duo v1 Reciprocal World Model

Ant Group's Technology Research Institute and Hangzhou-based AI company Westlake HeartChen have released and open-source…

3 Schritte1 BeiträgeAktualisiert 2026-11-22 01:53

Wie es dazu kam

  1. 2026-11-112026-11-11: The research team submitted the paper 'Duo: A Reciprocal World Model for Full-Duplex Interaction' to arXiv, introducing the RWM concept.arxiv.org ↗
  2. 2026-11-132026-11-13: 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. 2026-11-172026-11-17: Media outlets including Zhidx.com reported on the public release and open-source availability of the Duo v1 model.zhidx.com ↗

Was noch ungeklärt ist

  • 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?
Vollständiges Thema lesen →
Recherche fortsetzen →