AI Trends

Trending story Continuously updated

Ant Technology Research Institute and Westlake Xinchen Release and Open-Source Duo v1

2 reports1 discussions3 story updatesUpdated Nov 19 10:36

Understand this story

AI overview

On November 17, Ant Technology Research Institute and Westlake Xinchen released and open-sourced Duo v1, a bidirectional world model. Through collaboration between its Intent and Realization modules, the model enables virtual characters to continuously receive users’ audio and video input, autonomously decide whether to listen or respond, and generate corresponding audiovisual behaviors. It makes an interaction decision approximately every 0.7 seconds and is compressed into a four-step generator through three-stage training. In 180 checks across 30 real-world interactions, Duo v1 passed 168, for an overall pass rate of 93.3% and a content responsiveness rate of 73.3%. The technical report, training code, and model weights are now publicly available, and there have been no other factually substantiated developments.

AI-generated from public reports

Latest development · Report

Ant Technology Research Institute and Westlake Xinchen Release and Open-Source Bidirectional World Model Duo v1

Developments

1 confirmed developments

  1. Ant Technology Research Institute and Westlake Xinchen Release and Open-Source Duo v1

    Source:Phoenix Technology

Reports and discussion timeline

Browse reports and discussion by time. Their evidence levels remain distinct.

Nov 19

  1. VR Gyroscope

    Ant Technology Research Institute and Westlake Xinchen Release and Open-Source Bidirectional World Model Duo v1

    Ant Technology Research Institute and Westlake Xinchen released and open-sourced the bidirectional world model Duo v1 on November 17, enabling virtual characters to continuously perceive input and autonomously decide whether to speak or listen. The model uses a dual-module architecture comprising Intent and Realization, makes an interaction decision approximately every 0.7 seconds, and compresses audiovisual generation to four steps. Across 180 checks in 30 real-world interactions, Duo v1 passed 168, achieving an overall pass rate of 93.3% and a content responsiveness rate of 73.3%.

  2. English text is being prepared.Related discussion

    English text is being prepared. joined this discussion

    This post contributes to heat; it is not a formal report or confirmed fact.

Nov 18

  1. Phoenix Technology

    English text is being prepared.

    English text is being prepared.

Heat trend for this story

Current
12
Period peak
17 Nov 19 11:00
24-hour change
—

Move the pointer or select the chart to inspect hourly heat. Use the left and right arrow keys with a keyboard.

Related stories

  1. Related storiesMustafa Suleyman launches MAI Cyber 2 Flash and MAI Code 2
  2. Related storiesOpenAI launches GPT-6.1 Terra on ChatGPT Work and Codex
  3. Related storiesiFLYTEK releases iFlytek Spark X2.6 device-cloud integrated large model
  4. Related storiesMiniMax releases the M3.1 Flash model