diff --git a/tests/dfx/reliability/invalid_param_test/test_invalid_minicpmo_4_5_omni.py b/tests/dfx/reliability/invalid_param_test/test_invalid_minicpmo_4_5_omni.py
deleted file mode 100644
index 3bcd84b4a..000000000
--- a/tests/dfx/reliability/invalid_param_test/test_invalid_minicpmo_4_5_omni.py
+++ /dev/null
@@ -1,79 +0,0 @@
-# SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
-"""Invalid audio format rejection on MiniCPM-o 4.5: ``POST /v1/chat/completions``."""
-
-import os
-
-os.environ.setdefault("VLLM_WORKER_MULTIPROC_METHOD", "spawn")
-
-import pytest
-
-from tests.helpers.mark import hardware_test
-from tests.helpers.runtime import OmniServerParams, OpenAIClientHandler, dummy_messages_from_mix_data
-from tests.helpers.stage_config import get_deploy_config_path
-
-pytestmark = [pytest.mark.slow, pytest.mark.omni, pytest.mark.full_model]
-
-_MINICPMO_DEPLOY = get_deploy_config_path("minicpmo_4_5.yaml")
-
-_MINICPMO_SERVER = [
-    pytest.param(
-        OmniServerParams(
-            model="openbmb/MiniCPM-o-4_5",
-            stage_config_path=_MINICPMO_DEPLOY,
-            use_stage_cli=False,
-            server_args=[
-                "--trust-remote-code",
-            ],
-        ),
-        id="minicpmo_4_5",
-    )
-]
-
-
-def _system_prompt() -> dict[str, object]:
-    return {
-        "role": "system",
-        "content": [
-            {
-                "type": "text",
-                "text": (
-                    "You are MiniCPM-o 4.5, a virtual human capable of perceiving "
-                    "auditory and visual inputs, as well as generating text and speech."
-                ),
-            }
-        ],
-    }
-
-
-def _prompt() -> str:
-    return "What is the capital of China? Answer in 20 words."
-
-
-@hardware_test(res={"cuda": "H100", "npu": "A3"}, num_cards=1)
-@pytest.mark.parametrize("omni_server", _MINICPMO_SERVER, indirect=True)
-def test_invalid_audio_format_rejected(omni_server: OmniServerParams, openai_client: OpenAIClientHandler) -> None:
-    """
-    Test that invalid audio format is properly rejected.
-    Deploy Setting: default single GPU
-    Input Modal: text + invalid audio format
-    Output Modal: error
-    """
-    fake_base64 = "AAAA"  # not a real audio file
-
-    responses = openai_client.send_chat_completions_http_request(
-        {
-            "json": {
-                "model": omni_server.model,
-                "messages": dummy_messages_from_mix_data(
-                    system_prompt=_system_prompt(),
-                    audio_data_url=f"data:audio/invalid;base64,{fake_base64}",
-                    content_text=_prompt(),
-                ),
-                "stream": True,
-            },
-        },
-        err_code=400,
-        err_message="Invalid or unsupported audio file",
-    )
-    assert not responses[0].success
diff --git a/tests/dfx/reliability/invalid_param_test/test_invalid_omni_chat.py b/tests/dfx/reliability/invalid_param_test/test_invalid_omni_chat.py
index 0d026a37b..835d546e4 100644
--- a/tests/dfx/reliability/invalid_param_test/test_invalid_omni_chat.py
+++ b/tests/dfx/reliability/invalid_param_test/test_invalid_omni_chat.py
@@ -1,6 +1,6 @@
 # SPDX-License-Identifier: Apache-2.0
 # SPDX-FileCopyrightText: Copyright contributors to the vLLM-Omni project
-"""Invalid inputs on Qwen3-Omni: ``POST /v1/chat/completions``, ``WS /v1/video/chat/stream``, ``WS /v1/realtime``."""
+"""Invalid inputs on omni chat: ``POST /v1/chat/completions``, ``WS /v1/video/chat/stream``, ``WS /v1/realtime``."""
 
 from __future__ import annotations
 
@@ -10,7 +10,7 @@ from typing import Any
 import pytest
 
 from tests.helpers.mark import hardware_test
-from tests.helpers.runtime import OmniServer, OmniServerParams, OnlineOmniClient
+from tests.helpers.runtime import OmniServer, OmniServerParams, OnlineOmniClient, dummy_messages_from_mix_data
 from tests.helpers.stage_config import get_deploy_config_path
 
 pytestmark = [pytest.mark.slow, pytest.mark.omni]
@@ -75,6 +75,12 @@ def _chat_completions_request_without_expectations(omni_server: OmniServer, case
     elif case_id == "audio_format_unsupported":
         body["modalities"] = ["text", "audio"]
         body["audio"] = {"voice": "alloy", "format": "aac"}
+    elif case_id == "audio_data_url_corrupt":
+        body["messages"] = dummy_messages_from_mix_data(
+            audio_data_url="data:audio/invalid;base64,AAAA",
+            content_text="What is the capital of China? Answer in 20 words.",
+        )
+        body["stream"] = True
     else:
         raise AssertionError(f"unknown chat completions invalid case_id {case_id!r}")
     return {"json": body, "timeout": 120}
@@ -144,6 +150,12 @@ def _chat_completions_request_without_expectations(omni_server: OmniServer, case
             ("Invalid audio format", "aac", "Supported formats"),
             id="unsupported_audio_format",
         ),
+        pytest.param(
+            "audio_data_url_corrupt",
+            400,
+            "Invalid or corrupted audio data",
+            id="invalid_audio_data_url",
+        ),
     ],
 )
 @pytest.mark.parametrize("omni_server", _QWEN3_OMNI_SERVER, indirect=True)