diff --git a/vllm/entrypoints/cli/benchmark/main.py b/vllm/entrypoints/cli/benchmark/main.py
index 2ff98577c3634..48f34fce1d44c 100644
--- a/vllm/entrypoints/cli/benchmark/main.py
+++ b/vllm/entrypoints/cli/benchmark/main.py
@@ -32,6 +32,7 @@ class BenchmarkSubcommand(CLISubcommand):
     ) -> FlexibleArgumentParser:
         bench_parser = subparsers.add_parser(
             self.name,
+            help=self.help,
             description=self.help,
             usage=f"vllm {self.name} <bench_type> [options]",
         )
diff --git a/vllm/entrypoints/cli/serve.py b/vllm/entrypoints/cli/serve.py
index 96608f360e17b..77c7253aef06e 100644
--- a/vllm/entrypoints/cli/serve.py
+++ b/vllm/entrypoints/cli/serve.py
@@ -66,7 +66,11 @@ class ServeSubcommand(CLISubcommand):
         self, subparsers: argparse._SubParsersAction
     ) -> FlexibleArgumentParser:
         serve_parser = subparsers.add_parser(
-            self.name, description=DESCRIPTION, usage="vllm serve [model_tag] [options]"
+            self.name,
+            help="Launch a local OpenAI-compatible API server to serve LLM "
+            "completions via HTTP.",
+            description=DESCRIPTION,
+            usage="vllm serve [model_tag] [options]",
         )
 
         serve_parser = make_arg_parser(serve_parser)