Add comparison to previous model

README.md

@@ -45,6 +45,10 @@ Applied Claude-sourced datasets containing ~200k question/answer pairs for fine-
 <s>[INST]{prompt}[/INST]
 ```
 
+## Comparison
+
+In my non-exhaustive testing, this model performs as well or better as Llama3.1-8B-Sonnet in half the execution time.
+
 ## Release History
 
  * v0.1 - [2025-02-12] Initial release, trained to 1024 steps