diff --git a/tests/e2e/online_serving/test_omni_gen2_expansion.py b/tests/e2e/online_serving/test_omni_gen2_expansion.py
new file mode 100644
index 00000000000..d3901e667d7
--- /dev/null
+++ b/tests/e2e/online_serving/test_omni_gen2_expansion.py
@@ -0,0 +1,90 @@
+"""
+Online serving diffusion feature coverage for OmniGen2.
+
+Coverage:
+- Ulysses-SP
+- CFG-Parallel
+
+assert_diffusion_response validates successful generation and the expected
+image resolution.
+"""
+
+import pytest
+
+from tests.conftest import (
+    OmniServer,
+    OmniServerParams,
+    OpenAIClientHandler,
+    dummy_messages_from_mix_data,
+)
+from tests.utils import hardware_marks
+
+MODEL = "OmniGen2/OmniGen2"
+PROMPT = "A sleek concept car parked beside a glass building at sunrise, cinematic lighting, ultra detailed."
+NEGATIVE_PROMPT = "blurry, low quality, distorted, watermark"
+
+SINGLE_CARD_FEATURE_MARKS = hardware_marks(res={"cuda": "L4"})
+PARALLEL_FEATURE_MARKS = {
+    "usp": hardware_marks(res={"cuda": "L4"}, num_cards=3),
+    "cfg_parallel": hardware_marks(res={"cuda": "L4"}, num_cards=2),
+}
+
+
+def _get_omnigen2_feature_cases(model: str):
+    """Return OmniGen2 online serving cases for SP, and CFG-Parallel."""
+
+    return [
+        pytest.param(
+            OmniServerParams(
+                model=model,
+                server_args=[
+                    "--usp",
+                    "3",
+                ],
+            ),
+            id="ulysses_sp_3",
+            marks=PARALLEL_FEATURE_MARKS["usp"],
+        ),
+        pytest.param(
+            OmniServerParams(
+                model=model,
+                server_args=[
+                    "--cfg-parallel-size",
+                    "2",
+                ],
+            ),
+            id="cfg_parallel_2",
+            marks=PARALLEL_FEATURE_MARKS["cfg_parallel"],
+        ),
+    ]
+
+
+@pytest.mark.advanced_model
+@pytest.mark.diffusion
+@pytest.mark.parametrize(
+    "omni_server",
+    _get_omnigen2_feature_cases(MODEL),
+    indirect=True,
+)
+def test_omnigen2(
+    omni_server: OmniServer,
+    openai_client: OpenAIClientHandler,
+):
+    """Validate OmniGen2 online serving for SP, and CFG-Parallel."""
+
+    messages = dummy_messages_from_mix_data(content_text=PROMPT)
+
+    request_config = {
+        "model": omni_server.model,
+        "messages": messages,
+        "extra_body": {
+            "height": 1024,
+            "width": 1024,
+            "num_inference_steps": 2,
+            "negative_prompt": NEGATIVE_PROMPT,
+            "guidance_scale": 4.0,
+            "seed": 42,
+        },
+    }
+
+    openai_client.send_diffusion_request(request_config)