diff --git a/tests/benchmark/run.sh b/tests/benchmark/run.sh
new file mode 100755
index 0000000..d59b497
--- /dev/null
+++ b/tests/benchmark/run.sh
@@ -0,0 +1,119 @@
+#!/usr/bin/env bash
+set -e
+
+echo "# Running benchmarks..."
+echo ""
+
+BENCH_DIR="$(pwd)/tests/benchmark"
+IRC_BIN="$(pwd)/build/nix-irc"
+
+GREEN='\033[0;32m'
+BLUE='\033[0;34m'
+YELLOW='\033[0;33m'
+NC='\033[0m'
+
+get_ms() {
+	local time_str="$1"
+	if [[ $time_str =~ ([0-9]+)m([0-9.]+)s ]]; then
+		local mins="${BASH_REMATCH[1]}"
+		local secs="${BASH_REMATCH[2]}"
+		local ms
+		ms=$(awk "BEGIN {printf \"%.1f\", ($mins * 60000) + ($secs * 1000)}")
+		echo "$ms"
+	else
+		echo "0"
+	fi
+}
+
+run_benchmark() {
+	local name="$1"
+	local file="$2"
+
+	echo -e "${BLUE}=== $name ===${NC}"
+	echo ""
+
+	# Measure compilation time
+	echo -n "  Compilation time:     "
+	local compile_start
+	compile_start=$(date +%s%N)
+	"$IRC_BIN" "$file" /tmp/bench.nixir >/dev/null 2>&1
+	local compile_end
+	compile_end=$(date +%s%N)
+	local compile_ms=$(((compile_end - compile_start) / 1000000))
+	echo -e "${YELLOW}${compile_ms}ms${NC}"
+
+	# Source and IR sizes
+	local src_size
+	src_size=$(stat -c%s "$file" 2>/dev/null || stat -f%z "$file" 2>/dev/null)
+	local ir_size
+	ir_size=$(stat -c%s /tmp/bench.nixir 2>/dev/null || stat -f%z /tmp/bench.nixir 2>/dev/null)
+	local ratio=$((ir_size * 100 / src_size))
+	echo -e "  Source size:          ${src_size}B"
+	echo -e "  IR bundle size:       ${ir_size}B (${ratio}% of source)"
+
+	echo ""
+
+	# Native Nix evaluation (baseline)
+	echo -n "  Native Nix eval:      "
+	local native_total=0
+	for _ in {1..5}; do
+		local t
+		t=$( (time nix-instantiate --eval --strict "$file" >/dev/null 2>&1) 2>&1 | grep "real" | awk '{print $2}')
+		local ms
+		ms=$(get_ms "$t")
+		native_total=$(awk "BEGIN {print $native_total + $ms}")
+	done
+	local native_avg
+	native_avg=$(awk "BEGIN {printf \"%.1f\", $native_total / 5}")
+	echo -e "${GREEN}${native_avg}ms${NC} avg (5 runs)"
+
+	echo ""
+}
+
+echo "Measuring IR compilation speed and bundle size characteristics."
+echo ""
+
+run_benchmark "Simple Expression" "$BENCH_DIR/simple.nix"
+run_benchmark "Medium Complexity" "$BENCH_DIR/medium.nix"
+run_benchmark "Large Expression" "$BENCH_DIR/large.nix"
+
+# Overall statistics
+echo -e "${BLUE}=== Overall Statistics ===${NC}"
+echo ""
+
+testdir=$(mktemp -d)
+total_nix=0
+total_ir=0
+total_compile_time=0
+
+for f in "$BENCH_DIR"/*.nix; do
+	nixsize=$(stat -c%s "$f" 2>/dev/null || stat -f%z "$f" 2>/dev/null)
+	base=$(basename "$f" .nix)
+	irfile="${testdir}/${base}.nixir"
+
+	start=$(date +%s%N)
+	"$IRC_BIN" "$f" "$irfile" >/dev/null 2>&1
+	end=$(date +%s%N)
+	compile_time=$(((end - start) / 1000000))
+
+	if [ -f "$irfile" ]; then
+		irsize=$(stat -c%s "$irfile" 2>/dev/null || stat -f%z "$irfile" 2>/dev/null)
+		total_nix=$((total_nix + nixsize))
+		total_ir=$((total_ir + irsize))
+		total_compile_time=$((total_compile_time + compile_time))
+	fi
+done
+
+total_ratio=$((total_ir * 100 / total_nix))
+avg_compile_time=$((total_compile_time / 3))
+
+# TBH those are entirely unnecessary. However, I'm a sucker for data
+# and those are trivial to compile. Might as well. Who knows, maybe it'll
+# come in handy in the future.
+echo "  Total source size:     ${total_nix}B"
+echo "  Total IR size:         ${total_ir}B"
+echo "  Compression ratio:     ${total_ratio}% of source"
+echo "  Average compile time:  ${avg_compile_time}ms"
+echo ""
+
+rm -rf "$testdir"
diff --git a/tests/benchmark/run_benchmarks.sh b/tests/benchmark/run_benchmarks.sh
deleted file mode 100755
index 49c9ba8..0000000
--- a/tests/benchmark/run_benchmarks.sh
+++ /dev/null
@@ -1,69 +0,0 @@
-#!/usr/bin/env bash
-set -e
-
-echo ""
-
-PLUGIN_PATH="$(pwd)/build/nix-ir-plugin.so"
-BENCH_DIR="$(pwd)/tests/benchmark"
-IRC_BIN="$(pwd)/build/nix-irc"
-
-# Colors for output
-GREEN='\033[0;32m'
-BLUE='\033[0;34m'
-NC='\033[0m' # No Color
-
-run_benchmark() {
-	local name="$1"
-	local file="$2"
-
-	echo -e "${BLUE}Benchmark: $name${NC}"
-	echo "----------------------------------------"
-
-	# 1. Parse + Compile Time (nix-irc)
-	echo -n "  Parse + Compile: "
-	local compile_output=$( (time "$IRC_BIN" "$file" /tmp/bench.nixir 2>&1) 2>&1)
-	local compile_time=$(echo "$compile_output" | grep "real" | awk '{print $2}')
-	echo -e "${GREEN}$compile_time${NC}"
-
-	# 2. IR Size
-	if [ -f /tmp/bench.nixir ]; then
-		local ir_size=$(stat -f%z /tmp/bench.nixir 2>/dev/null || stat -c%s /tmp/bench.nixir 2>/dev/null)
-		echo -e "  IR Bundle Size:  ${GREEN}${ir_size} bytes${NC}"
-	fi
-
-	# 3. Native Nix evaluation time
-	echo -n "  Native Eval:     "
-	local native_time=$( (time nix-instantiate --eval --strict "$file" >/dev/null 2>&1) 2>&1 | grep "real" | awk '{print $2}')
-	echo -e "${GREEN}$native_time${NC}"
-
-	# 4. With Plugin evaluation time
-	echo -n "  Plugin Eval:     "
-	local plugin_time=$( (time nix-instantiate --plugin-files "$PLUGIN_PATH" --eval --strict "$file" >/dev/null 2>&1) 2>&1 | grep "real" | awk '{print $2}')
-	echo -e "${GREEN}$plugin_time${NC}"
-
-	echo ""
-}
-
-echo "# Running benchmarks..."
-echo ""
-
-run_benchmark "Simple Expressions" "$BENCH_DIR/simple.nix"
-run_benchmark "Medium Complexity" "$BENCH_DIR/medium.nix"
-run_benchmark "Large/Complex" "$BENCH_DIR/large.nix"
-
-# File size comparison
-echo -e "${BLUE}File Size Comparison${NC}"
-echo "----------------------------------------"
-testdir=$(mktemp -d)
-
-for f in "$BENCH_DIR"/*.nix; do
-	nixsize=$(stat -c%s "$f" 2>/dev/null || stat -f%z "$f" 2>/dev/null)
-	base=$(basename "$f" .nix)
-	irfile="${testdir}/${base}.nixir"
-	$IRC_BIN "$f" "$irfile" >/dev/null 2>&1
-	if [ -f "$irfile" ]; then
-		irsize=$(stat -c%s "$irfile" 2>/dev/null || stat -f%z "$irfile" 2>/dev/null)
-		ratio=$((irsize * 100 / nixsize))
-		echo "  $base: ${nixsize}B => ${irsize}B (${ratio}% of source)"
-	fi
-done