Core Insights - Yunzhisheng has launched the Unisound U1-OCR document intelligence foundational model, which claims to offer advantages such as SOTA performance, verifiability, plug-and-play usability, efficient deployment, and strong adaptability [2][5] - The model is designed to understand document layouts and extract deep semantic information, overcoming the limitations of traditional models that only read text and do not comprehend formatting [2][5] Performance Metrics - In the OmniDocBench V1.5 evaluation, the Unisound U1-OCR achieved a score of 95.1, marking it as a SOTA performer and surpassing other leading models such as GLM-OCR, Deepseek-OCR2, Gemini-3-Pro, and GPT-5.2 [2][5] - The model demonstrates a dual breakthrough in accuracy and generalization capabilities [2][5] Internal Testing Results - Internal business tests indicate that the model's information extraction and document classification capabilities exceed those of mainstream commercial and open-source models like Gemini-3-Flash and Qwen-235B-VL [2][5] - The Unisound U1-OCR shows particularly strong advantages in specialized business scenarios, such as medical admission records and discharge summaries, outperforming larger-scale general VLMs with its 3 billion parameter size [2][5]