Core Insights - The SuperCLUE platform released a new ranking of generative video models, with Wanxing Technology's Pangu-T2V model ranking fourth in China, surpassing OpenAI's Sora in overall score [1][6]. Group 1: Model Performance - Pangu-T2V achieved fourth place in the SuperCLUE-T2V benchmark across various categories, including general capability, video quality rendering, element construction, and dynamic shaping [1][5]. - The model is positioned in the "Excellence Leader" quadrant of the SuperCLUE-T2V video model quadrant, indicating its leading capabilities in general and application-specific performance [1][6]. Group 2: Technological Advancements - Wanxing Technology launched the Wanxing Tianmu Multimedia Model 2.0, which shows a 90% performance improvement over its predecessor, enhancing capabilities in professional camera work, multi-layered audio effects, and intelligent frame transitions [7][9]. - The new model is built on a vast dataset, enabling deep understanding across multiple modalities [7]. Group 3: Strategic Collaborations - Wanxing Technology has partnered with Huawei Cloud to establish an "AI Video Model Laboratory," aiming to explore innovations in large models and AI applications in digital creative scenarios [9][10]. - The collaboration leverages Wanxing's strengths in audio-visual models and Huawei's cloud infrastructure to drive technological and application innovations [10]. Group 4: Market Positioning - Wanxing Technology is recognized as a leading player in the digital creative software sector, with a global user base exceeding 1.5 billion across over 200 countries, often referred to as the "Chinese version of Adobe" [10]. - The company aims to continuously upgrade its model capabilities and provide innovative AI audio-visual solutions to enhance diverse creative scenarios [10].
文生视频大模型全球榜发布 万兴科技旗下万兴天幕2.0基模能力测评总分超Sora