diff --git a/benchmarks/MGET/dfly_bench/MGET_reproduce.md b/benchmarks/MGET/dfly_bench/MGET_reproduce.md
new file mode 100644
index 0000000..59b2b51
--- /dev/null
+++ b/benchmarks/MGET/dfly_bench/MGET_reproduce.md
@@ -0,0 +1,35 @@
+### Stateful setup:
+
+```
+dfbench setup --id maxqps2 --tune-network --ubuntu-version 24.04 \
+  --server-instance m7g.8xlarge  --server-arch arm64 \
+  --client-instance c6gn.8xlarge --client-arch arm64 \
+  --engines dragonfly,redis,valkey --availability-zone us-east-1c
+```
+
+### Test run:
+
+```
+dfbench run-commands-dfly \
+  --id maxqps2 --engines dragonfly,redis,valkey \
+  \
+  --command-template "MGET __key__" --command-name MGET \
+  --preload-template "SET __key__ __data__" \
+  --key-maximum 100000000 \
+  --test-time 60 --warmup-time 10 --trials 1 \
+  \
+  --dfly-bench-threads 32 --dfly-bench-conns 5 --dfly-bench-pipeline 30 \
+  --dfly-bench-data-size 128 --dfly-bench-key-dist U --dfly-bench-qps 0 \
+  \
+  --dragonfly_version s3://df-use1-pub/v2/dragonfly-aarch64.tar.gz \
+  --dragonfly_num_shards 16
+```
+
+### Expected results:
+
+  MGET
+  Engine      Throughput (median)   p50        p99        p99.9      Avg Latency
+  ──────      ───────────────────   ───        ───        ─────      ───────────
+  dragonfly   8.09M ops/s           0.549 ms   1.114 ms   1.432 ms   0.570 ms
+  redis       1.37M ops/s           3.408 ms   6.658 ms   7.939 ms   3.478 ms
+  valkey      1.49M ops/s           2.754 ms   8.701 ms   9.487 ms   3.211 ms
\ No newline at end of file
diff --git a/benchmarks/MSET/dfly_bench/MSET_reproduce.md b/benchmarks/MSET/dfly_bench/MSET_reproduce.md
new file mode 100644
index 0000000..4a9c67c
--- /dev/null
+++ b/benchmarks/MSET/dfly_bench/MSET_reproduce.md
@@ -0,0 +1,36 @@
+### Stateful setup:
+
+```
+dfbench setup --id maxqps2 --tune-network --ubuntu-version 24.04 \
+  --server-instance m7g.8xlarge  --server-arch arm64 \
+  --client-instance c6gn.8xlarge --client-arch arm64 \
+  --engines dragonfly,redis,valkey --availability-zone us-east-1c
+```
+
+### Test run:
+
+```
+dfbench run-commands-dfly \
+  --id maxqps2 --engines dragonfly,redis,valkey \
+  \
+  --command-template "MSET __key__ __data__" --command-name MSET --command-mutates \
+  --key-maximum 100000000 \
+  --test-time 300 --warmup-time 10 --trials 1 \
+  \
+  --dfly-bench-threads 32 --dfly-bench-conns 5 --dfly-bench-pipeline 30 \
+  --dfly-bench-data-size 128 --dfly-bench-key-dist U --dfly-bench-qps 0 \
+  \
+  --dragonfly_version s3://df-use1-pub/v2/dragonfly-aarch64.tar.gz \
+  --dragonfly_num_shards 16 --dragonfly_conn_use_incoming_cpu=true \
+  \
+  --redis_io_threads 8 --valkey_io_threads 8
+```
+
+### Expected results:
+
+  MSET
+  Engine      Throughput (median)   p50        p99         p99.9       Avg Latency
+  ──────      ───────────────────   ───        ───         ─────       ───────────
+  dragonfly   7.42M ops/s           0.519 ms   3.193 ms    10.310 ms   0.631 ms
+  redis       932.9K ops/s          4.390 ms   8.630 ms    15.157 ms   5.132 ms
+  valkey      1.24M ops/s           3.705 ms   11.262 ms   15.579 ms   3.843 ms
\ No newline at end of file
diff --git a/benchmarks/ZADD/dfly_bench/ZADD_reproduce.md b/benchmarks/ZADD/dfly_bench/ZADD_reproduce.md
index 5327627..07b026b 100644
--- a/benchmarks/ZADD/dfly_bench/ZADD_reproduce.md
+++ b/benchmarks/ZADD/dfly_bench/ZADD_reproduce.md
@@ -1,5 +1,15 @@
+### Stateful setup:
+
+```
+dfbench setup --id maxqps2 --tune-network --ubuntu-version 24.04 \
+  --server-instance m7g.8xlarge  --server-arch arm64 \
+  --client-instance c6gn.8xlarge --client-arch arm64 \
+  --engines dragonfly,redis,valkey --availability-zone us-east-1c
+```
+
 ### Test run:
 
+```
 dfbench run-commands-dfly \
   --id maxqps2 --engines dragonfly,redis,valkey \
   \
@@ -14,6 +24,7 @@ dfbench run-commands-dfly \
   --dragonfly_num_shards 16 --dragonfly_conn_use_incoming_cpu=true \
   \
   --redis_io_threads 8 --valkey_io_threads 8
+```
 
 ### Expected results:
 
diff --git a/benchmarks/ZSCORE/dfly_bench/ZSCORE_reproduce.md b/benchmarks/ZSCORE/dfly_bench/ZSCORE_reproduce.md
index 0a07b1d..996f08f 100644
--- a/benchmarks/ZSCORE/dfly_bench/ZSCORE_reproduce.md
+++ b/benchmarks/ZSCORE/dfly_bench/ZSCORE_reproduce.md
@@ -1,5 +1,15 @@
+### Stateful setup:
+
+```
+dfbench setup --id maxqps2 --tune-network --ubuntu-version 24.04 \
+  --server-instance m7g.8xlarge  --server-arch arm64 \
+  --client-instance c6gn.8xlarge --client-arch arm64 \
+  --engines dragonfly,redis,valkey --availability-zone us-east-1c
+```
+
 ### Test run:
 
+```
 dfbench run-commands-dfly \
   --id maxqps2 --engines dragonfly,redis,valkey \
   \
@@ -14,6 +24,7 @@ dfbench run-commands-dfly \
   \
   --dragonfly_version s3://df-use1-pub/v2/dragonfly-aarch64.tar.gz \
   --dragonfly_num_shards 16
+```
 
 ### Expected results: