PulseAugur
EN
LIVE 07:25:23
中文(ZH) 阿里视频大模型Wan3.0正式上线,行业评价“稳定、真实、有质感”

Alibaba launches Wan3.0 video generation model with enhanced realism

Alibaba has officially launched its video generation large model, Wan3.0, on August 24th. This latest iteration boasts significant upgrades in video generation length, creative versatility, reference capabilities, and realistic world rendering. Wan3.0 can now produce up to 30 seconds of video in a single generation and uniquely supports input formats like documents and spreadsheets, aiming for greater real-world accuracy. Early users and creators have praised the model for its stability, realism, and high-quality output, noting its ability to maintain consistency in characters, props, and narrative flow, thereby reducing the need for extensive post-generation adjustments. AI

IMPACT Enhances realism and stability in AI video generation, potentially accelerating adoption in media and advertising.

RANK_REASON Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=2 ai=1.0]

Read on 量子位 (QbitAI) →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

Alibaba launches Wan3.0 video generation model with enhanced realism

COVERAGE [2]

  1. 量子位 (QbitAI) TIER_1 中文(ZH) · 量子位的朋友们 ·

    Alibaba's large video model Wan3.0 officially launched, industry evaluation: "stable, realistic, textured"

    8月24日,阿里巴巴视频生成大模型Wan3.0正式上线。

  2. 雷峰网 (Leiphone) TIER_1 中文(ZH) ·

    Alibaba's large video model Wan3.0 officially launched, industry evaluation: "stable, realistic, textured"

    <p style="text-align: center;"><img src="https://static.leiphone.com/uploads/new/images/20260824/6a8bb4dae7d69.png?imageView2/2/w/740" /></p><p>8月24日,阿里巴巴视频生成大模型Wan3.0正式上线。该模型在生成时长、万能创作、全能参考以及真实世界还原等维度全面升级,单次可生成30秒视频,并首次支持doc、xls、ppt、pdf、md等文档格式输入,力求准确还原真实世界。公测以来,见证Wan模型家族历次更新的创作…