昆仑万维全新SkyReels正式焕新上线

Core Insights - Kunlun Wanwei has launched its AI video creation platform SkyReels, which is now available on both web and mobile applications, aimed at enabling global users to create professional-level content easily [1] Group 1: Product Features - SkyReels' core positioning is its one-stop and multi-modal capabilities, integrating top global AI models such as Google Veo 3.1, Sora 2, and others, offering functionalities like image generation, video generation, digital humans, and music generation [1][2] - The newly launched SkyReels V3 is based on Kunlun Wanwei's self-developed model, featuring a series of multi-modal video generation models that utilize a Multi-modal In Context Learning framework for pre-training and fine-tuning [1][2] - The platform introduces an "Agentic Copilot" mode, which includes a dual-core intelligent system supporting multi-modal input and output, catering to both immediate creative needs and in-depth professional tasks [2] Group 2: Technological Advancements - SkyReels V3 is the first in the industry to support multi-person, multi-turn dialogue with digital humans, allowing precise control over each character's speaking timing and rhythm, enhancing the natural flow of multi-character interactions [2] - The digital human functionality supports various scenarios, including film-level dialogue, e-commerce dual-host broadcasts, and game material creation, marking a significant advancement in audio-driven video generation capabilities [3] - Future developments in visual and audio generation models are expected to accelerate, with improvements in model effectiveness and controllability, while content generation costs are anticipated to decrease [3]