diff --git a/benchmarks/Makefile b/benchmarks/Makefile
index 51aa0942ab..cb66613105 100644
--- a/benchmarks/Makefile
+++ b/benchmarks/Makefile
@@ -48,6 +48,9 @@ TSBS_WORKERS ?= 4
 TSBS_QUERY_TYPE := cpu-max-all-8
 TSBS_DATA_FILE := /tmp/tsbs-data-$(TSBS_SCALE)-$(TSBS_START)-$(TSBS_END)-$(TSBS_STEP).gz
 TSBS_QUERY_FILE := /tmp/tsbs-queries-$(TSBS_QUERY_TYPE)-$(TSBS_SCALE)-$(TSBS_START)-$(TSBS_END)-$(TSBS_QUERIES).gz
+TSBS_LOAD_RESULT_CSV_FILE := /tmp/tsbs-load-$(TSBS_SCALE)-$(TSBS_START)-$(TSBS_END)-$(TSBS_STEP).csv
+TSBS_LOAD_RESULT_CSV_FILE_COMPARE ?=
+TSBS_PLOT_SCRIPT := $(shell pwd)/benchmarks/plot-load.sh
 # For cluster setup use http://vminsert:8480/insert/0/influx/write
 TSBS_WRITE_URLS ?= http://localhost:8428/write
 # For cluster setup use http://vmselect:8481/select/0/prometheus
@@ -76,7 +79,10 @@ tsbs-generate-data:
 
 # Load data into VictoriaMetrics
 tsbs-load-data:
-	cat $(TSBS_DATA_FILE) | gunzip | /tmp/tsbs/bin/tsbs_load_victoriametrics --workers=$(TSBS_WORKERS) --urls=$(TSBS_WRITE_URLS)
+	cat $(TSBS_DATA_FILE) | gunzip | /tmp/tsbs/bin/tsbs_load_victoriametrics \
+		--workers=$(TSBS_WORKERS) \
+		--urls=$(TSBS_WRITE_URLS) \
+		2>/dev/tty | tee $(TSBS_LOAD_RESULT_CSV_FILE)
 	curl -s $(TSBS_METRICS_URL) | grep \
 		-e process_cpu_seconds_user_total \
 		-e process_cpu_seconds_system_total \
@@ -133,3 +139,16 @@ tsbs-run-queries-all:
 
 	# Too heavy: retrieves 1B samples
 	# $(MAKE) tsbs-run-queries TSBS_QUERY_TYPE=double-groupby-all
+
+# Plot the data load `per interval metrics/s` measurements
+#
+# To plot measurements collected during the recent benchmark, run
+# make tsbs-plot-load
+#
+# To plot the measurements from some other benchmark, run
+# make tsbs-plot-load TSBS_LOAD_RESULT_CSV_FILE=/path/to/file.csv
+#
+# To plot the measurements from two benchmarks, run
+# make tsbs-plot-load TSBS_LOAD_RESULT_CSV_FILE=/path/to/file1.csv TSBS_LOAD_RESULT_CSV_FILE_COMPARE=/path/to/file2.csv
+tsbs-plot-load:
+	$(TSBS_PLOT_SCRIPT) $(TSBS_LOAD_RESULT_CSV_FILE) $(TSBS_LOAD_RESULT_CSV_FILE_COMPARE)
diff --git a/benchmarks/plot-load.sh b/benchmarks/plot-load.sh
new file mode 100755
index 0000000000..2692978f3a
--- /dev/null
+++ b/benchmarks/plot-load.sh
@@ -0,0 +1,55 @@
+#!/bin/bash
+
+# Check the number of arguments
+if [ "$#" -lt 1 ] || [ "$#" -gt 2 ]; then
+  echo "Usage: $(basename $0) file1.csv [file2.csv]"
+  exit 1
+fi
+
+file1="$1"
+file2="$2"
+
+# Check if the files exist
+if [ ! -f "$file1" ]; then
+  echo "File not found: $file1"
+  exit 1
+fi
+
+if [ -n "$file2" ] && [ ! -f "$file2" ]; then
+  echo "File not found: $file2"
+  exit 1
+fi
+
+# Temporary file for plot data
+plotdata="/tmp/tsbs_plot.dat"
+
+if [ -z "$file2" ]; then
+  # === One file ===
+  awk -F, '/^Summary:/ {exit} NR > 1 {print $1, $2}' "$file1" > "$plotdata"
+else
+  # === Two files ===
+
+  # Get time and per.metric/s from file1
+  awk -F, '/^Summary:/ {exit} NR > 1 {print $1, $2}' "$file1" > /tmp/file1.dat
+  # Get per.metric/s from file2
+  awk -F, '/^Summary:/ {exit} NR > 1 {print $2}' "$file2" > /tmp/file2.dat
+  # Merge by rows: time, val1, val2
+  paste /tmp/file1.dat /tmp/file2.dat | awk '{print $1, $2, $3}' > "$plotdata"
+fi
+
+# === Build plot dynamically ===
+gnuplot -persist <<-EOF
+  set datafile separator " "
+  set title "per.metric/s"
+  set xlabel "Timestamp"
+  set xdata time
+  set timefmt "%s"
+  set format x "%H:%M:%S"
+  set ylabel "per. metric/s"
+  set format y "%'.0f"
+  set grid
+
+  plot "$plotdata" using 1:2 with lines title "$(basename "$file1")" \
+  $( [[ -n "$file2" ]] && echo ", \\
+        \"$plotdata\" using 1:3 with lines title \"$(basename "$file2")\"" )
+EOF