Skywork AI Enhances Deep Research Agent with Multimodal Capabilities

Skywork AI has launched the upgraded Skywork Deep Research Agent v2, featuring enhanced multimodal capabilities and improved efficiency, announced in a press release. This upgrade is part of Skywork AI's Technology Release Week, which began on August 11, 2025.

The Skywork Deep Research Agent v2 introduces a "Multi-Modal Deep Research" feature, integrating multi-modal retrieval, understanding, and generation. This allows the agent to analyze mixed text-and-image formats, such as financial report graphs and social media visuals, enhancing decision-making insights and output quality. The agent employs technologies like MM-Crawler and asynchronous parallel processing to achieve simultaneous text reading and image comprehension.

Additionally, the "Multi-Modal Deep Browser Agent" redefines social media analytics by offering advanced multimodal comprehension and automated data analysis. It can extract insights from images and videos on platforms like Twitter and Instagram, transforming raw data into visually digestible reports. The agent is currently in alpha testing, with a full release expected soon.

Skywork Deep Research Agent v2 also achieves state-of-the-art performance in complex task execution, including advanced information retrieval and synthesis. It has set new benchmarks in accuracy, particularly in its proprietary Parallel Thinking mode, demonstrating exceptional scalability and potential.