diff --git a/tensorflow/lite/tools/benchmark/benchmark_tflite_model.cc b/tensorflow/lite/tools/benchmark/benchmark_tflite_model.cc
index 197907ec9e8..2edbbd06ec4 100644
--- a/tensorflow/lite/tools/benchmark/benchmark_tflite_model.cc
+++ b/tensorflow/lite/tools/benchmark/benchmark_tflite_model.cc
@@ -573,12 +573,8 @@ TfLiteStatus BenchmarkTfLiteModel::Init() {
   // Install profilers if necessary right after interpreter is created so that
   // any memory allocations inside the TFLite runtime could be recorded if the
   // installed profiler profile memory usage information.
-  if (params_.Get<bool>("enable_op_profiling")) {
-    profiling_listener_.reset(new ProfilingListener(
-        interpreter_.get(),
-        params_.Get<int32_t>("max_profiling_buffer_entries")));
-    AddListener(profiling_listener_.get());
-  }
+  profiling_listener_ = MayCreateProfilingListener();
+  if (profiling_listener_) AddListener(profiling_listener_.get());
 
   interpreter_->UseNNAPI(params_.Get<bool>("use_legacy_nnapi"));
   interpreter_->SetAllowFp16PrecisionForFp32(params_.Get<bool>("allow_fp16"));
@@ -771,6 +767,14 @@ std::unique_ptr<tflite::OpResolver> BenchmarkTfLiteModel::GetOpResolver()
   return std::unique_ptr<tflite::OpResolver>(resolver);
 }
 
+std::unique_ptr<BenchmarkListener>
+BenchmarkTfLiteModel::MayCreateProfilingListener() const {
+  if (!params_.Get<bool>("enable_op_profiling")) return nullptr;
+  return std::unique_ptr<BenchmarkListener>(new ProfilingListener(
+      interpreter_.get(),
+      params_.Get<int32_t>("max_profiling_buffer_entries")));
+}
+
 TfLiteStatus BenchmarkTfLiteModel::RunImpl() { return interpreter_->Invoke(); }
 
 }  // namespace benchmark
diff --git a/tensorflow/lite/tools/benchmark/benchmark_tflite_model.h b/tensorflow/lite/tools/benchmark/benchmark_tflite_model.h
index a6fc38a6180..3778cc968bd 100644
--- a/tensorflow/lite/tools/benchmark/benchmark_tflite_model.h
+++ b/tensorflow/lite/tools/benchmark/benchmark_tflite_model.h
@@ -71,6 +71,10 @@ class BenchmarkTfLiteModel : public BenchmarkModel {
   // Allow subclasses to create a customized Op resolver during init.
   virtual std::unique_ptr<tflite::OpResolver> GetOpResolver() const;
 
+  // Create a BenchmarkListener that's specifically for TFLite profiling if
+  // necessary.
+  virtual std::unique_ptr<BenchmarkListener> MayCreateProfilingListener() const;
+
   void CleanUp();
 
   std::unique_ptr<tflite::FlatBufferModel> model_;
@@ -103,8 +107,8 @@ class BenchmarkTfLiteModel : public BenchmarkModel {
 
   std::vector<InputLayerInfo> inputs_;
   std::vector<InputTensorData> inputs_data_;
-  std::unique_ptr<BenchmarkListener> profiling_listener_;
-  std::unique_ptr<BenchmarkListener> gemmlowp_profiling_listener_;
+  std::unique_ptr<BenchmarkListener> profiling_listener_ = nullptr;
+  std::unique_ptr<BenchmarkListener> gemmlowp_profiling_listener_ = nullptr;
   TfLiteDelegatePtrMap delegates_;
 
   std::mt19937 random_engine_;