视频大模型

Search documents
一举登顶!生数科技新模型Vidu Q1,用极致效果坐稳性价比之王
Jiang Nan Shi Bao· 2025-04-23 07:03
Core Insights - ShengShu Technology has launched its new video model Vidu Q1 globally, which has garnered significant attention in the industry due to its performance enhancements and new features [1] - Vidu Q1 has achieved top rankings in the VBench-1.0 and VBench-2.0 assessments, surpassing leading models such as Runway, OpenAI Sora, and Kuaishou's Kling [1][6] Performance Metrics - In the VBench-1.0 rankings, Vidu Q1 outperformed competitors in video quality and semantic consistency, achieving state-of-the-art (SOTA) levels in various dimensions [9] - The model supports the generation of 5-second, 1080P high-quality videos, marking a significant advancement in AI video commercialization [9][11] Features and Upgrades - Vidu Q1 offers a comprehensive upgrade in "extreme high quality," including: - Cinema-grade HD quality with 1080P output for both narrative and detailed character expressions [11] - Enhanced frame capabilities allowing for masterful camera movements with just two images [11] - Improved animation styles with high dynamic motion performance [11] - AI sound effect generation that allows for precise audio creation and layering [11] - The model is accessible via web and app platforms without the need for an invitation, promoting widespread usability at a competitive cost of only 0.3 yuan per second [11]