AI News (2026/2/3): GLM-OCR Released: SOTA Performance for Complex Documents
Executive Summary:
Zhipu AI has officially released and open-sourced the professional-level OCR model GLM-OCR, achieving "small size, high precision" document parsing capabilities with just 0.9B parameters. The model topped the authoritative OmniDocBench V1.5 evaluation with a score of 94.6, demonstrating state-of-the-art (SOTA) performance in tasks such as text recognition, formula recognition, table parsing, and information extraction. It is particularly optimized for challenging scenarios like handwritten text, complex tables, and code documents, supporting batch processing of PDFs and images with a throughput of 1.86 pages per second.
GLM-OCR Released: SOTA Performance for Complex Documents
News Details
Zhipu AI officially released and open-sourced its latest OCR model, GLM-OCR, on Tuesday, February 3, 2023. The model achieves outstanding document parsing capabilities with just 0.9B parameters, particularly excelling in high-difficulty scenarios such as handwritten text, complex tables, and code documents. GLM-OCR scored an impressive 94.6 points in the authoritative OmniDocBench V1.5 evaluation, making it one of the best-performing OCR models currently available.
Key Highlights
【Small Size, High Precision】: The GLM-OCR model contains only 0.9B parameters but achieves SOTA performance in multiple tasks. This feature allows the model to run efficiently on resource-constrained devices, reducing deployment costs and complexity.
【SOTA Performance in Multiple Tasks】: GLM-OCR scored 94.6 points in the OmniDocBench V1.5 evaluation, particularly excelling in text recognition, formula recognition, table parsing, and information extraction. High-precision parsing in these tasks is crucial for fields such as academic research, financial analysis, and legal document processing.
【Optimized for Challenging Scenarios】: The model is specifically optimized for challenging scenarios like handwritten text, complex tables, and code documents. Improved accuracy in handwritten text recognition makes the digitization of historical documents and notes more convenient; enhanced complex table parsing capabilities improve the efficiency of handling financial statements and research data; and support for code documents provides software developers with a new tool for document management.
AI-ALL In-Depth Commentary
The release of GLM-OCR marks another significant breakthrough for Zhipu AI in the field of OCR technology. Its "small size, high precision" characteristics not only meet the needs of resource-constrained devices but also provide more efficient and accurate solutions for various applications. Especially in fields such as academic research, financial analysis, and legal document processing, the performance of GLM-OCR will significantly enhance work efficiency and data accuracy. Additionally, open-sourcing this model will further drive technological innovation and application expansion within the community, injecting new vitality into the development of the AI ecosystem.