update readme

README.md

@@ -1,13 +1,17 @@
 # Real5-OmniDocBench
 
-<div align="center">
-[[📜 arXiv]](https://arxiv.org/pdf/2603.04205) | [[Dataset (🤗Hugging Face)]](https://huggingface.co/datasets/PaddlePaddle/Real5-OmniDocBench)
-</div>
-
 **Real5-OmniDocBench** is a brand-new benchmark oriented toward real-world scenarios, which we constructed based on the OmniDocBench v1.5 dataset. The dataset comprises five distinct scenarios: Scanning, Warping, Screen-Photography, Illumination, and Skew. Apart from the Scanning category, all images were manually acquired via handheld mobile devices to closely simulate real-world conditions. Each subset maintains a one-to-one correspondence with the original OmniDocBench, strictly adhering to its ground-truth annotations and evaluation protocols. Given its empirical and realistic nature, this dataset serves as a rigorous benchmark for assessing the robustness of document parsing models in practical applications. 
 
 ---
 
+## Updates
+
+- [2026/03/05] Release paper and update mertics
+  - The paper has been released on [arXiv](https://arxiv.org/abs/2512.03069).
+  - Update DeepSeek-OCR 2, GLM-OCR model evaluation.
+
+---
+
 ## Key Features
 
 ### 1. Real-world Scenarios
@@ -1330,4 +1334,8 @@ If you use Real5-OmniDocBench in your research, please cite our dataset paper an
       primaryClass={cs.CV},
       url={https://arxiv.org/abs/2603.04205}, 
 }
-```
+```
+
+## Links
+
+- Paper: [Real5-OmniDocBench](https://arxiv.org/pdf/2603.04205)