Model Releases

Wan-Streamer v0.2 released with 550ms end-to-end response latency and 640×368 real-time video dialogue

Original title:Wan-Streamer v0.2:响应延迟仅550ms,让 AI 真正与你"面对面"

AI summary

Tongyi Lab released Wan-Streamer v0.2, featuring 550ms end-to-end response latency (200ms model latency plus 350ms network latency) and raising output resolution from 192×336 in v0.1 to 640×368 at 25FPS.

Key Takeaways

  • Tongyi Lab released Wan-Streamer v0.
  • 2,featuring 550ms end-to-end response latency (200ms model latency plus 350ms network latency) and…
  • 1 to 640×368 at 25FPS.

Quick Facts

Published
Updated
Source
公众号:通义实验室(千问)
Category
Model Releases
AI score
76
Read the original Read the full article at the original source
Why it matters

550ms端到端延迟把实时视频对话从拼接流水线推入单模型交互循环,原先因响应太慢而无法沉浸的口语陪练与面试模拟,开始具备自然对话的交互基础。

Source & copyright

This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.

Original article · AIHOT entry

How to Cite

When citing this article, keep the title, date, and permalink, and credit Orange News as the source.

Orange News: Wan-Streamer v0.2 released with 550ms end-to-end response latency and 640×368 real-time video dialogue. July 17, 2026. https://new.nat6.net/en/news/cmsdu3gxy014urofz1avi6jfh/

View all →