diff --git a/tools/compare.py b/tools/compare.py
index 66eed932c2..01d2c89f50 100755
--- a/tools/compare.py
+++ b/tools/compare.py
@@ -238,10 +238,10 @@ def main():
         options_contender = ['--benchmark_filter=%s' % filter_contender]
 
     # Run the benchmarks and report the results
-    json1 = json1_orig = gbench.util.run_or_load_benchmark(
-        test_baseline, benchmark_options + options_baseline)
-    json2 = json2_orig = gbench.util.run_or_load_benchmark(
-        test_contender, benchmark_options + options_contender)
+    json1 = json1_orig = gbench.util.sort_benchmark_results(gbench.util.run_or_load_benchmark(
+        test_baseline, benchmark_options + options_baseline))
+    json2 = json2_orig = gbench.util.sort_benchmark_results(gbench.util.run_or_load_benchmark(
+        test_contender, benchmark_options + options_contender))
 
     # Now, filter the benchmarks so that the difference report can work
     if filter_baseline and filter_contender:
diff --git a/tools/gbench/Inputs/test4_run.json b/tools/gbench/Inputs/test4_run.json
new file mode 100644
index 0000000000..eaa005f3a9
--- /dev/null
+++ b/tools/gbench/Inputs/test4_run.json
@@ -0,0 +1,96 @@
+{
+  "benchmarks": [
+    {
+      "name": "99 family 0 instance 0 repetition 0",
+      "run_type": "iteration",
+      "family_index": 0,
+      "per_family_instance_index": 0,
+      "repetition_index": 0
+    },
+    {
+      "name": "98 family 0 instance 0 repetition 1",
+      "run_type": "iteration",
+      "family_index": 0,
+      "per_family_instance_index": 0,
+      "repetition_index": 1
+    },
+    {
+      "name": "97 family 0 instance 0 aggregate",
+      "run_type": "aggregate",
+      "family_index": 0,
+      "per_family_instance_index": 0,
+      "aggregate_name": "9 aggregate"
+    },
+
+
+    {
+      "name": "96 family 0 instance 1 repetition 0",
+      "run_type": "iteration",
+      "family_index": 0,
+      "per_family_instance_index": 1,
+      "repetition_index": 0
+    },
+    {
+      "name": "95 family 0 instance 1 repetition 1",
+      "run_type": "iteration",
+      "family_index": 0,
+      "per_family_instance_index": 1,
+      "repetition_index": 1
+    },
+    {
+      "name": "94 family 0 instance 1 aggregate",
+      "run_type": "aggregate",
+      "family_index": 0,
+      "per_family_instance_index": 1,
+      "aggregate_name": "9 aggregate"
+    },
+
+
+
+
+    {
+      "name": "93 family 1 instance 0 repetition 0",
+      "run_type": "iteration",
+      "family_index": 1,
+      "per_family_instance_index": 0,
+      "repetition_index": 0
+    },
+    {
+      "name": "92 family 1 instance 0 repetition 1",
+      "run_type": "iteration",
+      "family_index": 1,
+      "per_family_instance_index": 0,
+      "repetition_index": 1
+    },
+    {
+      "name": "91 family 1 instance 0 aggregate",
+      "run_type": "aggregate",
+      "family_index": 1,
+      "per_family_instance_index": 0,
+      "aggregate_name": "9 aggregate"
+    },
+
+
+    {
+      "name": "90 family 1 instance 1 repetition 0",
+      "run_type": "iteration",
+      "family_index": 1,
+      "per_family_instance_index": 1,
+      "repetition_index": 0
+    },
+    {
+      "name": "89 family 1 instance 1 repetition 1",
+      "run_type": "iteration",
+      "family_index": 1,
+      "per_family_instance_index": 1,
+      "repetition_index": 1
+    },
+    {
+      "name": "88 family 1 instance 1 aggregate",
+      "run_type": "aggregate",
+      "family_index": 1,
+      "per_family_instance_index": 1,
+      "aggregate_name": "9 aggregate"
+    }
+  ]
+}
diff --git a/tools/gbench/report.py b/tools/gbench/report.py
index 69b643e92b..6bea82f6bf 100644
--- a/tools/gbench/report.py
+++ b/tools/gbench/report.py
@@ -1,9 +1,11 @@
-import unittest
 """report.py - Utilities for reporting statistics about benchmark results
 """
+
+import unittest
 import os
 import re
 import copy
+import random
 
 from scipy.stats import mannwhitneyu
 
@@ -912,6 +914,49 @@ def test_json_diff_report(self):
             assert_measurements(self, out, expected)
 
 
+class TestReportSorting(unittest.TestCase):
+    @classmethod
+    def setUpClass(cls):
+        def load_result():
+            import json
+            testInputs = os.path.join(
+                os.path.dirname(
+                    os.path.realpath(__file__)),
+                'Inputs')
+            testOutput = os.path.join(testInputs, 'test4_run.json')
+            with open(testOutput, 'r') as f:
+                json = json.load(f)
+            return json
+
+        cls.json = load_result()
+
+    def test_json_diff_report_pretty_printing(self):
+        import util
+
+        expected_names = [
+            "99 family 0 instance 0 repetition 0",
+            "98 family 0 instance 0 repetition 1",
+            "97 family 0 instance 0 aggregate",
+            "96 family 0 instance 1 repetition 0",
+            "95 family 0 instance 1 repetition 1",
+            "94 family 0 instance 1 aggregate",
+            "93 family 1 instance 0 repetition 0",
+            "92 family 1 instance 0 repetition 1",
+            "91 family 1 instance 0 aggregate",
+            "90 family 1 instance 1 repetition 0",
+            "89 family 1 instance 1 repetition 1",
+            "88 family 1 instance 1 aggregate"
+        ]
+
+        for n in range(len(self.json['benchmarks']) ** 2):
+            random.shuffle(self.json['benchmarks'])
+            sorted_benchmarks = util.sort_benchmark_results(self.json)[
+                'benchmarks']
+            self.assertEqual(len(expected_names), len(sorted_benchmarks))
+            for out, expected in zip(sorted_benchmarks, expected_names):
+                self.assertEqual(out['name'], expected)
+
+
 def assert_utest(unittest_instance, lhs, rhs):
     if lhs['utest']:
         unittest_instance.assertAlmostEqual(
diff --git a/tools/gbench/util.py b/tools/gbench/util.py
index 661c4bad8d..5d0012c0cb 100644
--- a/tools/gbench/util.py
+++ b/tools/gbench/util.py
@@ -5,6 +5,7 @@
 import tempfile
 import subprocess
 import sys
+import functools
 
 # Input file type enumeration
 IT_Invalid = 0
@@ -119,6 +120,23 @@ def load_benchmark_results(fname):
         return json.load(f)
 
 
+def sort_benchmark_results(result):
+    benchmarks = result['benchmarks']
+
+    # From inner key to the outer key!
+    benchmarks = sorted(
+        benchmarks, key=lambda benchmark: benchmark['repetition_index'] if 'repetition_index' in benchmark else -1)
+    benchmarks = sorted(
+        benchmarks, key=lambda benchmark: 1 if 'run_type' in benchmark and benchmark['run_type'] == "aggregate" else 0)
+    benchmarks = sorted(
+        benchmarks, key=lambda benchmark: benchmark['per_family_instance_index'] if 'per_family_instance_index' in benchmark else -1)
+    benchmarks = sorted(
+        benchmarks, key=lambda benchmark: benchmark['family_index'] if 'family_index' in benchmark else -1)
+
+    result['benchmarks'] = benchmarks
+    return result
+
+
 def run_benchmark(exe_name, benchmark_flags):
     """
     Run a benchmark specified by 'exe_name' with the specified