diff --git a/benchmark.py b/benchmark.py
new file mode 100644
index 0000000..ea0908a
--- /dev/null
+++ b/benchmark.py
@@ -0,0 +1,24 @@
+from proofow import pow256 as proofow
+from time import perf_counter
+
+
+def benchmark_pow(difficulty, iterations):
+    values = []
+    for i in range(iterations):
+        pw = proofow(difficulty)
+        values.append(measure(pw.work))
+    return sum(values) / len(values)
+
+
+def measure(what):
+    start = perf_counter()
+    what()
+    return perf_counter() - start
+
+
+if __name__ == '__main__':
+    ITER = 1000
+    print('Benchmarking proofow:\n')
+    for i in range(5):
+        print('{} iterations of difficuluty {} takes {} units of time'.format(ITER, i, benchmark_pow(i, ITER)*1000000))
+    print('\nYou should see that each difficulity level roughly doubles in terms of computation time.')