Update README.md

README.md

@@ -34,7 +34,7 @@ For question-answering, Jan-v1 shows a significant performance gain from model s
 
 These benchmarks evaluate the model's conversational and instructional capabilities.
 
-![image/png](https://cdn-uploads.huggingface.co/production/uploads/65713d70f56f9538679e5a56/Dqtzn-3HP0ftIo8x124Ca.png)
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/65713d70f56f9538679e5a56/f3bzNYRuA_iTFQIcvu6Rr.png)
 
 ## Quick Start