diff --git a/src/c++/perf_analyzer/README.md b/src/c++/perf_analyzer/README.md
index 881447651..806a3db8f 100644
--- a/src/c++/perf_analyzer/README.md
+++ b/src/c++/perf_analyzer/README.md
@@ -73,6 +73,9 @@ changes in performance as you experiment with different optimization strategies.
   [TorchServe](docs/benchmarking.md#benchmarking-torchserve) can be used as the
   inference server in addition to the default Triton server
 
+- [LLMs](docs/llm.md) can also be measured and charcterized with specific metrics
+  like token-to-token latency
+
 <br>
 
 # Quick Start