diff --git a/vllm_omni/entrypoints/cli/logo.py b/vllm_omni/entrypoints/cli/logo.py
new file mode 100644
index 000000000..addfe3593
--- /dev/null
+++ b/vllm_omni/entrypoints/cli/logo.py
@@ -0,0 +1,45 @@
+import re
+
+from vllm.logger import current_formatter_type, init_logger
+
+logger = init_logger(__name__)
+
+ORANGE = "\033[38;5;208m"
+BLUE = "\033[34m"
+WHITE = "\033[97m"
+PURPLE = "\033[35m"
+RESET = "\033[0m"
+
+# Official vLLM Logo Base
+VLLM_L1 = f"       {WHITE}█     █     █▄   ▄█{RESET}"
+VLLM_L2 = f" {ORANGE}▄▄{RESET} {BLUE}▄█{RESET} {WHITE}█     █     █ ▀▄▀ █{RESET}"
+VLLM_L3 = f"  {ORANGE}█{RESET}{BLUE}▄█▀{RESET} {WHITE}█     █     █     █{RESET}"
+VLLM_L4 = f"   {BLUE}▀▀{RESET}  {WHITE}▀▀▀▀▀ ▀▀▀▀▀ ▀     ▀{RESET}"
+
+GAP_L1 = "    "
+GAP_L2 = f"{WHITE}  ▄▄▄ {RESET}"
+GAP_L3 = f"{WHITE}      {RESET}"
+GAP_L4 = "      "
+
+O_L1 = f"{BLUE}   ▄▀▀{ORANGE}▀▀▄ {RESET}"
+O_L2 = f"{BLUE} █    {ORANGE}█ {RESET}"
+O_L3 = f"{BLUE} █    {ORANGE}█ {RESET}"
+O_L4 = f"{PURPLE}  ▀▀▀▀  {RESET}"
+MNI_L1 = f"{WHITE}█▄   ▄█ █▄    █ ▀█▀ {RESET}"
+MNI_L2 = f"{WHITE}█ ▀▄▀ █ █ ▀▄  █  █  {RESET}"
+MNI_L3 = f"{WHITE}█     █ █   ▀▄█  █  {RESET}"
+MNI_L4 = f"{WHITE}▀     ▀ ▀     ▀ ▀▀▀ {RESET}"
+
+LOGO = f"""{VLLM_L1}{GAP_L1}{O_L1}{MNI_L1}
+{VLLM_L2}{GAP_L2}{O_L2}{MNI_L2}
+{VLLM_L3}{GAP_L3}{O_L3}{MNI_L3}
+{VLLM_L4}{GAP_L4}{O_L4}{MNI_L4}
+"""
+
+
+_ANSI_RE = re.compile(r"\033\[[0-9;]*m")
+
+
+def log_logo() -> None:
+    logo = LOGO if current_formatter_type(logger) == "color" else _ANSI_RE.sub("", LOGO)
+    logger.info(logo)
diff --git a/vllm_omni/entrypoints/cli/serve.py b/vllm_omni/entrypoints/cli/serve.py
index a59cc5673..e28637f6a 100644
--- a/vllm_omni/entrypoints/cli/serve.py
+++ b/vllm_omni/entrypoints/cli/serve.py
@@ -6,6 +6,7 @@ diffusion models (e.g., Qwen-Image) through the same CLI interface.
 """
 
 import argparse
+import os
 
 import uvloop
 from vllm.entrypoints.cli.types import CLISubcommand
@@ -14,6 +15,7 @@ from vllm.entrypoints.utils import VLLM_SUBCMD_PARSER_EPILOG
 from vllm.logger import init_logger
 from vllm.utils.argparse_utils import FlexibleArgumentParser
 
+from vllm_omni.entrypoints.cli.logo import log_logo
 from vllm_omni.entrypoints.openai.api_server import omni_run_server
 
 logger = init_logger(__name__)
@@ -45,6 +47,10 @@ class OmniServeCommand(CLISubcommand):
 
     @staticmethod
     def cmd(args: argparse.Namespace) -> None:
+        if not os.environ.get("VLLM_DISABLE_LOG_LOGO"):
+            os.environ["VLLM_DISABLE_LOG_LOGO"] = "1"
+            log_logo()
+
         # If model is specified in CLI (as positional arg), it takes precedence
         if hasattr(args, "model_tag") and args.model_tag is not None:
             args.model = args.model_tag