Update README.md

README.md

@@ -16,16 +16,20 @@ tags:
 # ActPRM
 
 ## Links
-- [GitHub](https://github.com/sail-sg/ActivePRM)
-- [ActivePRM Collection](https://huggingface.co/collections/sail/active-prm-67fccbc10c94c41a6d6b26d9)
+- 💻 [GitHub](https://github.com/sail-sg/ActivePRM)
+- 🤗 [ActivePRM Collection](https://huggingface.co/collections/sail/active-prm-67fccbc10c94c41a6d6b26d9)
 
 ## Introduction
 
-We release our process reward model effectively trained with active learning based on `Qwen/Qwen2.5-Math-7B-Instruct`.
+We release our process reward model effectively trained with active learning based on `Qwen/Qwen2.5-Math-PRM-7B`.
+The model achieved 76.0% on [ProcessBench](https://github.com/QwenLM/ProcessBench) and 66.7% on [PRMBench](https://github.com/ssmisya/PRMBench).
+Both achieve new state-of-the-art (SOTA) performance among 7B parameter models (as of 15/04/2025).
+
+ <p align="center">
+  <img src="https://raw.githubusercontent.com/sail-sg/ActivePRM/refs/heads/master/assets/processbench.png" width=80%/>
+  <img src="https://raw.githubusercontent.com/sail-sg/ActivePRM/refs/heads/master/assets/prmbench.png" width=80%/>
+</p>
 
- <!-- <p align="center">
-  <img src="https://raw.githubusercontent.com/sail-sg/ActivePRM/refs/heads/main/assets/figure1.png" width=70%/>
-</p> -->
 
 ### Usage
 We provide a code snippet to show the usage of our PRM.