Update README.md

README.md

@@ -18,7 +18,7 @@ Designed for **low-latency inference** on **NVIDIA RTX GPUs**, leveraging:
 - Quantization (W4A16)
 - Custom vLLM generation via `embedl-models`
 
-FlashHead matches the baseline **Llama-3.2-3B-Instruct** within rounding on standard evaluations (MMLU-Pro, HellaSwag, GSM8K, etc.) and, in combination with quantization, achieves **H200-level latency** on **RTX Ada** GPUs.
+FlashHead matches the Llama-3.2-3B-Instruct baseline within rounding error on common benchmarks (MMLU-Pro, HellaSwag, GSM8K, etc.) and, combined with quantization, delivers SOTA on-device latency
 
 ---