Update README.md

README.md

@@ -68,6 +68,7 @@ The following table summarizes the performance of various quantized models (**Q4
 | Qwen3 4b Instruct 2507         | 50.87%         |
 | Qwen3 4b                       | 50.43%         |
 | Qwen2.5 3b Instruct            | 47.94%         |
+| Ministral 3 3b Instruct 2512   | 43.65%         |
 | Gemma 3n E2B                   | 42.78%         |
 | Llama 3.2 3b Instruct          | 41.73%         |
 | Gemma 3 4b Instruct            | 41.52%         |