Update README.md

README.md

@@ -26,9 +26,11 @@ tags:
 
 Tipping the scales at 101 layers and 1215 tensors... the monster lives.
 
-Stronger, more compact coding with enhanced understanding of your instructions.
+Two monsters in fact.
 
-A fully overpowered CODING ENGINE based on:
+Each model generates stronger, more compact code with an enhanced understanding of your instructions and follows them to the letter.
+
+These overpowered CODING ENGINEs are based on:
 
 "Qwen2.5-Coder-32B-Instruct" 
 
@@ -40,11 +42,13 @@ and
 
 [ https://huggingface.co/open-r1/OlympicCoder-32B ]
 
-In one MASSSIVE 51B merge - that is greater in performance than the sum of both models.
+In one MASSSIVE 51B merge that is stronger in performance than both models.
 
 In two configs... too.
 
-Quants Q4_K_S (two - one of each config) uploading... these are unaltered quants for primary testing.
+Quants Q4_K_S (two - one of each config) are available. 
+
+These are unaltered quants for primary testing.
 
 CONFIGS:
 - #1 -> Qwen2.5-Coder-32B-Instruct primary/start, with OlympicCoder-32B as "finalizer".
@@ -70,7 +74,7 @@ Requirements:
 - rep pen: 1.1 (can be lower)
 - Jinja Template (embedded) or CHATML template.
 
-Refer to either "Qwen2.5-Coder-32B-Instruct" and/or "OlympicCoder-32B" for additional settings, benchmarks and usage.
+Refer to either "Qwen2.5-Coder-32B-Instruct" and/or "OlympicCoder-32B" repos (above) for additional settings, benchmarks and usage.
 
 ---