diff --git a/.buildkite/common/ci_source_file_dependencies.yml b/.buildkite/common/ci_source_file_dependencies.yml
index 99bbffa7548..edf0a480216 100644
--- a/.buildkite/common/ci_source_file_dependencies.yml
+++ b/.buildkite/common/ci_source_file_dependencies.yml
@@ -317,7 +317,7 @@ source_file_dependencies:
 
   diffusion_wan22_perf:
     - *wan22
-    - tests/dfx/perf/scripts/run_diffusion_benchmark.py
+    - tests/dfx/perf/scripts/run_benchmark.py
     - tests/dfx/perf/tests/test_wan22_i2v_vllm_omni.json
 
   diffusion_wan22_reliability:
@@ -343,7 +343,7 @@ source_file_dependencies:
 
   diffusion_cosmos3_perf:
     - *cosmos3
-    - tests/dfx/perf/scripts/run_diffusion_benchmark.py
+    - tests/dfx/perf/scripts/run_benchmark.py
     - tests/dfx/perf/tests/test_cosmos3_vllm_omni.json
 
   diffusion_qwen_image_function:
@@ -416,7 +416,7 @@ source_file_dependencies:
 
   diffusion_minimax_h3_perf:
     - *minimax_h3
-    - tests/dfx/perf/scripts/run_diffusion_benchmark.py
+    - tests/dfx/perf/scripts/run_benchmark.py
     - tests/dfx/perf/tests/test_minimax_h3_vllm_omni.json
 
   diffusion_hunyuan_video_function:
@@ -430,7 +430,7 @@ source_file_dependencies:
 
   diffusion_hunyuan_video_perf:
     - *hunyuan_video
-    - tests/dfx/perf/scripts/run_diffusion_benchmark.py
+    - tests/dfx/perf/scripts/run_benchmark.py
     - tests/dfx/perf/tests/test_hunyuanvideo15_t2v_vllm_omni.json
     - tests/dfx/perf/tests/test_hunyuanvideo15_i2v_vllm_omni.json
 
@@ -457,6 +457,11 @@ source_file_dependencies:
     - tests/e2e/online_serving/test_lingbot_video_moe.py
     - tests/e2e/offline_inference/test_lingbot_world_v2.py
 
+  diffusion_lingbot_perf:
+    - *lingbot_video
+    - tests/dfx/perf/scripts/run_benchmark.py
+    - tests/dfx/perf/tests/test_lingbot_video_vllm_omni.json
+
 
   # Invalid-param jobs: model business code only, plus the reliability scripts.
   reliability_invalid_param_h100:
diff --git a/.buildkite/cuda/test-nightly.yml b/.buildkite/cuda/test-nightly.yml
index 843b8d9a339..c69dcb14662 100644
--- a/.buildkite/cuda/test-nightly.yml
+++ b/.buildkite/cuda/test-nightly.yml
@@ -53,80 +53,62 @@ steps:
       - label: ":full_moon: Omni · Accuracy Test"
         source_file_dependencies: omni_qwen3_omni_accuracy
         timeout_in_minutes: 180
+        artifact_paths:
+          - tests/e2e/accuracy/qwen3_omni/results/qwen_omni_acc/*.json
         commands:
           - export SEED_TTS_WER_EVAL=1
           - export SEED_TTS_EVAL_DEVICE=cuda:1
-          - |
-            set +e
-            pytest -s -v tests/e2e/accuracy/qwen3_omni/test_qwen3_omni.py -m "full_model and H100 and B200 and cards_2" --run-level full_model
-            EXIT=$$?
-            buildkite-agent artifact upload "tests/e2e/accuracy/qwen3_omni/results/qwen_omni_acc/*.json"
-            exit $$EXIT
+          - pytest -s -v tests/e2e/accuracy/qwen3_omni/test_qwen3_omni.py -m "full_model and H100 and B200 and cards_2" --run-level full_model
 
       - label: ":full_moon: Omni · MiniCPM-o 4.5 · Accuracy Test"
         source_file_dependencies: omni_minicpmo_4_5_accuracy
         timeout_in_minutes: 180
+        artifact_paths:
+          - tests/e2e/accuracy/minicpmo_4_5/results/*.json
         commands:
           - export SEED_TTS_WER_EVAL=1
           - export SEED_TTS_EVAL_DEVICE=cuda:1
-          - |
-            set +e
-            pytest -s -v tests/e2e/accuracy/minicpmo_4_5/test_minicpmo_4_5.py -m "full_model and H100 and B200 and cuda and (cards_1 or cards_2)" --run-level full_model
-            EXIT=$$?
-            buildkite-agent artifact upload "tests/e2e/accuracy/minicpmo_4_5/results/*.json"
-            exit $$EXIT
+          - pytest -s -v tests/e2e/accuracy/minicpmo_4_5/test_minicpmo_4_5.py -m "full_model and H100 and B200 and cuda and (cards_1 or cards_2)" --run-level full_model
 
       - label: ":full_moon: Omni · Perf Test · No Async Chunk"
         source_file_dependencies: omni_qwen3_omni_perf
         key: nightly-omni-performance-no-async-chunk
         timeout_in_minutes: 300
+        artifact_paths:
+          - tests/dfx/perf/results/*.json
         commands:
           - export BENCHMARK_DIR=tests/dfx/perf/results
-          - |
-            set +e
-            pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_qwen3_omni_no_async_chunk.json -m "H100 and B200 and cards_2"
-            EXIT=$$?
-            buildkite-agent artifact upload "tests/dfx/perf/results/*.json"
-            exit $$EXIT
+          - pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_qwen3_omni_no_async_chunk.json -m "H100 and B200 and cards_2"
 
       - label: ":full_moon: Omni · Perf Test · Async Chunk"
         source_file_dependencies: omni_qwen3_omni_perf
         key: nightly-omni-performance-async-chunk
         timeout_in_minutes: 180
+        artifact_paths:
+          - tests/dfx/perf/results/*.json
         commands:
           - export BENCHMARK_DIR=tests/dfx/perf/results
-          - |
-            set +e
-            pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_qwen3_omni_async_chunk.json -m "H100 and B200 and full_model and cards_2"
-            EXIT=$$?
-            buildkite-agent artifact upload "tests/dfx/perf/results/*.json"
-            exit $$EXIT
+          - pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_qwen3_omni_async_chunk.json -m "H100 and B200 and full_model and cards_2"
 
       - label: ":full_moon: Omni · MiniCPM-o 4.5 · Perf Test"
         source_file_dependencies: omni_minicpmo_4_5_perf
         key: nightly-omni-performance-minicpmo-4-5
         timeout_in_minutes: 180
+        artifact_paths:
+          - tests/dfx/perf/results/*.json
         commands:
           - export BENCHMARK_DIR=tests/dfx/perf/results
-          - |
-            set +e
-            pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_minicpmo_4_5.json -m "H100 and B200 and cards_1"
-            EXIT=$$?
-            buildkite-agent artifact upload "tests/dfx/perf/results/*.json"
-            exit $$EXIT
+          - pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_minicpmo_4_5.json -m "H100 and B200 and cards_1"
 
       - label: ":full_moon: Omni · MiniCPM-o 4.5 · Duplex Seed-TTS Perf Test"
         source_file_dependencies: omni_minicpmo_4_5_duplex_perf
         key: nightly-omni-performance-minicpmo-4-5-duplex-seed-tts
         timeout_in_minutes: 180
+        artifact_paths:
+          - tests/dfx/perf/results/*.json
         commands:
           - export BENCHMARK_DIR=tests/dfx/perf/results
-          - |
-            set +e
-            pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_minicpmo_4_5_duplex_seed_tts.json -m "H100 and B200 and cards_1"
-            EXIT=$$?
-            buildkite-agent artifact upload "tests/dfx/perf/results/*.json"
-            exit $$EXIT
+          - pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_minicpmo_4_5_duplex_seed_tts.json -m "H100 and B200 and cards_1"
 
       - label: ":full_moon: Omni · Multi-Replica Startup Test with 4x H100"
         source_file_dependencies: omni_qwen3_omni_function
@@ -153,29 +135,23 @@ steps:
         source_file_dependencies: tts_qwen3_tts_perf
         key: nightly-tts-performance-single
         timeout_in_minutes: 180
+        artifact_paths:
+          - tests/dfx/perf/results/*.json
         commands:
           - export BENCHMARK_DIR=tests/dfx/perf/results
           - export VLLM_ALLOW_LONG_MAX_MODEL_LEN="1"
-          - |
-            set +e
-            pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_tts.json -m "H100 and B200 and cards_1"
-            EXIT=$$?
-            buildkite-agent artifact upload "tests/dfx/perf/results/*.json"
-            exit $$EXIT
+          - pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_tts.json -m "H100 and B200 and cards_1"
 
       - label: ":full_moon: TTS · Perf Test · 2-GPU"
         source_file_dependencies: tts_qwen3_tts_perf
         key: nightly-tts-performance-2gpu
         timeout_in_minutes: 180
+        artifact_paths:
+          - tests/dfx/perf/results/*.json
         commands:
           - export BENCHMARK_DIR=tests/dfx/perf/results
           - export VLLM_ALLOW_LONG_MAX_MODEL_LEN="1"
-          - |
-            set +e
-            pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_tts.json -m "H100 and B200 and cards_2"
-            EXIT=$$?
-            buildkite-agent artifact upload "tests/dfx/perf/results/*.json"
-            exit $$EXIT
+          - pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_tts.json -m "H100 and B200 and cards_2"
 
   - group: ":card_index_dividers: Tiny Model Tests [Multi-GPU]"
     key: nightly-tiny-model-test-group
@@ -277,33 +253,27 @@ steps:
         source_file_dependencies: diffusion_qwen_image_perf
         key: nightly-diffusion-x2iat-performance-qwen-image-single
         timeout_in_minutes: 180
+        artifact_paths:
+          - tests/dfx/perf/results/diffusion_result_*.json
+          - tests/dfx/perf/results/logs/*.log
         commands:
           - export DIFFUSION_BENCHMARK_DIR=tests/dfx/perf/results
           - export CACHE_DIT_VERSION=1.5.0
-          - |
-            set +e
-            pytest -s -v tests/dfx/perf/scripts/run_diffusion_benchmark.py --test-config-file tests/dfx/perf/tests/test_qwen_image_vllm_omni.json -m "H100 and B200 and cards_1"
-            EXIT=$$?
-            buildkite-agent artifact upload "tests/dfx/perf/results/diffusion_result_*.json"
-            buildkite-agent artifact upload "tests/dfx/perf/results/logs/*.log"
-            exit $$EXIT
+          - pytest -s -v tests/dfx/perf/scripts/run_diffusion_benchmark.py --test-config-file tests/dfx/perf/tests/test_qwen_image_vllm_omni.json -m "H100 and B200 and cards_1"
 
       - label: ":full_moon: Diffusion X2I(&A&T) · Perf Test · Qwen-Image · 4-GPU"
         source_file_dependencies: diffusion_qwen_image_perf
         key: nightly-diffusion-x2iat-performance-qwen-image-4gpu
         timeout_in_minutes: 180
+        artifact_paths:
+          - tests/dfx/perf/results/diffusion_result_*.json
+          - tests/dfx/perf/results/logs/*.log
         commands:
           - export DIFFUSION_BENCHMARK_DIR=tests/dfx/perf/results
           - export CACHE_DIT_VERSION=1.5.0
           # Do not pin DIFFUSION_ATTENTION_BACKEND: H100 auto-selects FLASH_ATTN;
           # B200 rejects explicit FLASH_ATTN without FA4 and uses CUDNN/TRTLLM.
-          - |
-            set +e
-            pytest -s -v tests/dfx/perf/scripts/run_diffusion_benchmark.py --test-config-file tests/dfx/perf/tests/test_qwen_image_vllm_omni.json -m "H100 and B200 and cards_4"
-            EXIT=$$?
-            buildkite-agent artifact upload "tests/dfx/perf/results/diffusion_result_*.json"
-            buildkite-agent artifact upload "tests/dfx/perf/results/logs/*.log"
-            exit $$EXIT
+          - pytest -s -v tests/dfx/perf/scripts/run_diffusion_benchmark.py --test-config-file tests/dfx/perf/tests/test_qwen_image_vllm_omni.json -m "H100 and B200 and cards_4"
 
   - group: ":card_index_dividers: Diffusion X2V Model Test"
     key: nightly-diffusion-x2v-group
@@ -399,15 +369,11 @@ steps:
         source_file_dependencies: diffusion_wan22_perf
         key: nightly-diffusion-x2v-performance-single
         timeout_in_minutes: 180
+        artifact_paths:
+          - tests/dfx/perf/results/*.json
         commands:
-          - export DIFFUSION_BENCHMARK_DIR=tests/dfx/perf/results
-          - |
-            set +e
-            pytest -s -v tests/dfx/perf/scripts/run_diffusion_benchmark.py --test-config-file tests/dfx/perf/tests/test_wan22_i2v_vllm_omni.json -m "H100 and B200 and cards_1"
-            EXIT=$$?
-            buildkite-agent artifact upload "tests/dfx/perf/results/diffusion_result_*.json"
-            buildkite-agent artifact upload "tests/dfx/perf/results/logs/*.log"
-            exit $$EXIT
+          - export BENCHMARK_DIR=tests/dfx/perf/results
+          - pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_wan22_i2v_vllm_omni.json -m "H100 and B200 and cards_1"
 
       - label: ":full_moon: Diffusion X2V · Perf Test · 2-GPU"
         source_file_dependencies:
@@ -415,16 +381,16 @@ steps:
           - diffusion_cosmos3_perf
         key: nightly-diffusion-x2v-performance-2gpu
         timeout_in_minutes: 180
+        artifact_paths:
+          - tests/dfx/perf/results/*.json
         commands:
-          - export DIFFUSION_BENCHMARK_DIR=tests/dfx/perf/results
+          - export BENCHMARK_DIR=tests/dfx/perf/results
           - |
             set +e
-            pytest -s -v tests/dfx/perf/scripts/run_diffusion_benchmark.py --test-config-file tests/dfx/perf/tests/test_wan22_i2v_vllm_omni.json -m "H100 and B200 and cards_2"
+            pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_wan22_i2v_vllm_omni.json -m "H100 and B200 and cards_2"
             EXIT1=$$?
-            pytest -s -v tests/dfx/perf/scripts/run_diffusion_benchmark.py --test-config-file tests/dfx/perf/tests/test_cosmos3_vllm_omni.json -m "H100 and B200 and cards_2"
+            pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_cosmos3_vllm_omni.json -m "H100 and B200 and cards_2"
             EXIT2=$$?
-            buildkite-agent artifact upload "tests/dfx/perf/results/diffusion_result_*.json"
-            buildkite-agent artifact upload "tests/dfx/perf/results/logs/*.log"
             if [ $$EXIT1 -ne 0 ]; then exit $$EXIT1; fi
             exit $$EXIT2
 
@@ -432,16 +398,16 @@ steps:
         source_file_dependencies: diffusion_hunyuan_video_perf
         key: nightly-diffusion-x2v-performance-hunyuanvideo15
         timeout_in_minutes: 180
+        artifact_paths:
+          - tests/dfx/perf/results/*.json
         commands:
-          - export DIFFUSION_BENCHMARK_DIR=tests/dfx/perf/results
+          - export BENCHMARK_DIR=tests/dfx/perf/results
           - |
             set +e
-            pytest -s -v tests/dfx/perf/scripts/run_diffusion_benchmark.py --test-config-file tests/dfx/perf/tests/test_hunyuanvideo15_t2v_vllm_omni.json -m "H100 and B200 and (cards_1 or cards_2)"
+            pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_hunyuanvideo15_t2v_vllm_omni.json -m "H100 and B200 and (cards_1 or cards_2)"
             EXIT1=$$?
-            pytest -s -v tests/dfx/perf/scripts/run_diffusion_benchmark.py --test-config-file tests/dfx/perf/tests/test_hunyuanvideo15_i2v_vllm_omni.json -m "H100 and B200 and (cards_1 or cards_2)"
+            pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_hunyuanvideo15_i2v_vllm_omni.json -m "H100 and B200 and (cards_1 or cards_2)"
             EXIT2=$$?
-            buildkite-agent artifact upload "tests/dfx/perf/results/diffusion_result_*.json"
-            buildkite-agent artifact upload "tests/dfx/perf/results/logs/*.log"
             if [ $$EXIT1 -ne 0 ]; then exit $$EXIT1; fi
             exit $$EXIT2
 
@@ -449,15 +415,21 @@ steps:
         source_file_dependencies: diffusion_minimax_h3_perf
         key: nightly-diffusion-x2v-performance-minimax-h3
         timeout_in_minutes: 180
+        artifact_paths:
+          - tests/dfx/perf/results/*.json
         commands:
-          - export DIFFUSION_BENCHMARK_DIR=tests/dfx/perf/results
-          - |
-            set +e
-            pytest -s -v tests/dfx/perf/scripts/run_diffusion_benchmark.py --test-config-file tests/dfx/perf/tests/test_minimax_h3_vllm_omni.json -m "H100 and B200 and cards_4"
-            EXIT=$$?
-            buildkite-agent artifact upload "tests/dfx/perf/results/diffusion_result_*.json"
-            buildkite-agent artifact upload "tests/dfx/perf/results/logs/*.log"
-            exit $$EXIT
+          - export BENCHMARK_DIR=tests/dfx/perf/results
+          - pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_minimax_h3_vllm_omni.json -m "H100 and B200 and cards_4"
+
+      - label: ":full_moon: Diffusion X2V · Perf Test · LingBot"
+        source_file_dependencies: diffusion_lingbot_perf
+        key: nightly-diffusion-x2v-performance-lingbot
+        timeout_in_minutes: 180
+        artifact_paths:
+          - tests/dfx/perf/results/*.json
+        commands:
+          - export BENCHMARK_DIR=tests/dfx/perf/results
+          - pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_lingbot_video_vllm_omni.json -m "H100 and B200 and cards_1"
 
   - group: ":card_index_dividers: Diffusion Test"
     key: nightly-diffusion-test-group
@@ -533,6 +505,9 @@ steps:
       - nightly-diffusion-x2v-performance-single
       - nightly-diffusion-x2v-performance-2gpu
     if: build.env("NIGHTLY") == "1" && build.env("EMAIL_DISTRIBUTION") == "1"
+    artifact_paths:
+      - tests/dfx/perf/results/*.xlsx
+      - tests/dfx/perf/results/*.html
     commands:
       - pip install openpyxl
       - export DEFAULT_INPUT_DIR=tests/dfx/perf/results
@@ -548,7 +523,5 @@ steps:
       - python tools/nightly/generate_nightly_perf_excel.py
       - python tools/nightly/generate_nightly_perf_html.py
       - python tools/nightly/send_nightly_email.py --report-file "tests/dfx/perf/results/*.xlsx, tests/dfx/perf/results/*.html"
-      - buildkite-agent artifact upload "tests/dfx/perf/results/*.xlsx"
-      - buildkite-agent artifact upload "tests/dfx/perf/results/*.html"
     agents:
       queue: "cpu_queue_premerge"
diff --git a/.buildkite/cuda/test-weekly.yml b/.buildkite/cuda/test-weekly.yml
index 70ab5935673..5c193945e34 100644
--- a/.buildkite/cuda/test-weekly.yml
+++ b/.buildkite/cuda/test-weekly.yml
@@ -12,6 +12,8 @@ steps:
     depends_on: upload-weekly-pipeline
     if: build.env("WEEKLY") == "1"
     timeout_in_minutes: 60
+    artifact_paths:
+      - coverage-core-model-cpu-*.xml.gz
     commands:
       - |
         set +e
@@ -19,7 +21,6 @@ steps:
         pytest -sv tests/ -m 'core_model and cpu' --cov=vllm_omni --cov-report=term-missing:skip-covered --cov-report=xml:$${REPORT}
         EXIT=$$?
         gzip -9 -f "$${REPORT}"
-        buildkite-agent artifact upload "$${REPORT}.gz"
         exit $$EXIT
     mirror_hardwares: l4_1
 
@@ -79,41 +80,32 @@ steps:
         source_file_dependencies: omni_qwen3_omni_perf
         key: weekly-omni-performance-vllm-text
         timeout_in_minutes: 180
+        artifact_paths:
+          - tests/dfx/perf/results/*.json
         commands:
           - export BENCHMARK_DIR=tests/dfx/perf/results
-          - |
-            set +e
-            pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_qwen3_omni_vllm_text.json
-            EXIT=$$?
-            buildkite-agent artifact upload "tests/dfx/perf/results/*.json"
-            exit $$EXIT
+          - pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_qwen3_omni_vllm_text.json
         mirror_hardwares: h100_2
 
       - label: ":full_moon: Omni · Perf Test · Async Chunk · Random"
         key: weekly-omni-performance-async-chunk-random
         timeout_in_minutes: 180
+        artifact_paths:
+          - tests/dfx/perf/results/*.json
         commands:
           - export BENCHMARK_DIR=tests/dfx/perf/results
-          - |
-            set +e
-            pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_qwen3_omni_async_chunk.json -m "H100 and slow"
-            EXIT=$$?
-            buildkite-agent artifact upload "tests/dfx/perf/results/*.json"
-            exit $$EXIT
+          - pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_qwen3_omni_async_chunk.json -m "H100 and slow"
         mirror_hardwares: h100_2
 
       - label: ":full_moon: Omni · Perf Test · Multi-Replica"
         source_file_dependencies: omni_qwen3_omni_perf
         key: weekly-omni-performance-multi-replicas
         timeout_in_minutes: 180
+        artifact_paths:
+          - tests/dfx/perf/results/*.json
         commands:
           - export BENCHMARK_DIR=tests/dfx/perf/results
-          - |
-            set +e
-            pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_qwen3_omni_multi_replicas.json
-            EXIT=$$?
-            buildkite-agent artifact upload "tests/dfx/perf/results/*.json"
-            exit $$EXIT
+          - pytest -s -v tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_qwen3_omni_multi_replicas.json
         mirror_hardwares: h100_3
 
   - group: ":card_index_dividers: E2E Tests"
diff --git a/.buildkite/npu/test-npu-nightly.yml b/.buildkite/npu/test-npu-nightly.yml
index 0c39a3f93a1..601a7bf4bd4 100644
--- a/.buildkite/npu/test-npu-nightly.yml
+++ b/.buildkite/npu/test-npu-nightly.yml
@@ -286,11 +286,10 @@ steps:
           HF_TOKEN: "${HF_TOKEN}"
           DIFFUSION_ATTENTION_BACKEND: TORCH_SDPA
         commands:
-          - export DIFFUSION_BENCHMARK_DIR=tests/dfx/perf/results
+          - export BENCHMARK_DIR=tests/dfx/perf/results
           - |
             set +e
-            pytest -s -v tests/dfx/perf/scripts/run_diffusion_benchmark.py -m npu --test-config-file tests/dfx/perf/tests/test_wan22_i2v_vllm_omni.json
+            pytest -s -v tests/dfx/perf/scripts/run_benchmark.py -m npu --test-config-file tests/dfx/perf/tests/test_wan22_i2v_vllm_omni.json
             EXIT=$$?
-            buildkite-agent artifact upload "tests/dfx/perf/results/diffusion_result_*.json"
-            buildkite-agent artifact upload "tests/dfx/perf/results/logs/*.log"
+            buildkite-agent artifact upload "tests/dfx/perf/results/*.json"
             exit $$EXIT
diff --git a/docs/contributing/ci/test_examples/l4_performance_tests.inc.md b/docs/contributing/ci/test_examples/l4_performance_tests.inc.md
index feb4f38ede7..981e6089e3c 100644
--- a/docs/contributing/ci/test_examples/l4_performance_tests.inc.md
+++ b/docs/contributing/ci/test_examples/l4_performance_tests.inc.md
@@ -7,16 +7,17 @@ When you want to add L4-level ***performance test*** cases, add entries to JSON
 | Omni (nightly) | `run_benchmark.py` | `test_qwen3_omni_no_async_chunk.json`, `test_qwen3_omni_async_chunk.json` (`full_model` without `slow` in `mark`) |
 | Omni (weekly) | `run_benchmark.py` | `test_qwen3_omni_async_chunk.json` (CUDA only), `test_qwen3_omni_vllm_text.json`, `test_qwen3_omni_multi_replicas.json` (`slow` in `mark`; **Perf Test** in `test-weekly.yml`) |
 | TTS | `run_benchmark.py` | `test_tts.json`, `test_voxcpm2.json`, `test_higgs_audio_v3.json` |
-| Diffusion | `run_diffusion_benchmark.py` | `test_qwen_image_vllm_omni.json`, `test_bagel_vllm_omni.json`, `test_wan22_i2v_vllm_omni.json`, `test_cosmos3_vllm_omni.json`, … |
+| Diffusion (`/v1/chat/completions`) | `run_diffusion_benchmark.py` | `test_qwen_image_vllm_omni.json`, `test_bagel_vllm_omni.json`, … |
+| Diffusion (`/v1/images/generations`, `/v1/images/edits`, `/v1/videos`) | `run_benchmark.py` | `test_wan22_i2v_vllm_omni.json`, `test_cosmos3_vllm_omni.json`, `test_lingbot_video_vllm_omni.json`, … |
 
 #### How runners pick cases
 
 Without **`--test-config-file`**, each runner scans all `*.json` under `tests/dfx/perf/tests/` but only keeps its own model type:
 
-- **`run_benchmark.py`**: omni and TTS cases only (skips diffusion JSON).
-- **`run_diffusion_benchmark.py`**: diffusion cases only (skips omni / TTS JSON).
+- **`run_benchmark.py`**: cases whose `benchmark_params` use ``dataset_name`` (omni, TTS, and image/video OpenAI generation via `vllm bench serve --omni`).
+- **`run_diffusion_benchmark.py`**: cases whose `benchmark_params` use ``dataset`` (diffusion client schema, including custom jsonl).
 
-Diffusion cases are detected when the JSON has `server_type` (typically `"vllm-omni"`) or `"diffusion"` in the `mark` array. Omni / TTS JSON has neither.
+Diffusion-script cases are detected by `is_diffusion_perf_config()`: presence of ``dataset`` (without ``dataset_name``) in `benchmark_params`. Marks / `server_type` / endpoint values are not used for this split.
 
 #### Running perf cases
 
@@ -60,16 +61,16 @@ Pass **`--test-config-file`** to load one JSON file, or omit it for the bulk sca
 
 ##### Overview
 
-| Field              | Required       | Description                                  |
-| ------------------ | -------------- | -------------------------------------------- |
-| test_name          | Yes            | Unique identifier for the test case          |
-| mark               | No             | Pytest marks; see **`mark` field** below     |
-| server_params      | Yes            | Server-side configuration parameters         |
-| benchmark_params   | Yes            | Benchmark running parameters                 |
-| server_type        | Diffusion only | Routes case to run_diffusion_benchmark.py    |
-| benchmark_endpoint | Diffusion only | Benchmark API path                           |
+| Field              | Required  | Description                                                                              |
+| ------------------ | --------- | ---------------------------------------------------------------------------------------- |
+| test_name          | Yes       | Unique identifier for the test case                                                      |
+| mark               | No        | Pytest marks; see **`mark` field** below                                                 |
+| server_params      | Yes       | Server-side configuration parameters                                                     |
+| benchmark_params   | Yes       | Benchmark running parameters                                                             |
+| server_type        | Diffusion | Only for diffusion-script JSON; omit on omni-bench generation cases                      |
+| benchmark_endpoint | Optional  | Legacy diffusion custom-jsonl alias; prefer `benchmark_params[].endpoint`                |
 
-Omit `mark` only for configs not meant to be filtered by `-m`. `server_type` is typically `"vllm-omni"`. `benchmark_endpoint` examples: `/v1/videos`, `/v1/images/generations`.
+Omit `mark` only for configs not meant to be filtered by `-m`. Cases that call `/v1/images/edits`, `/v1/images/generations`, or `/v1/videos` use the same `benchmark_params` schema as Omni/TTS (`dataset_name`, `endpoint`, `extra_body`) and are executed by `run_benchmark.py` — do not set `server_type` or `task` on those cases. Remaining diffusion cases (usually `/v1/chat/completions`, or custom jsonl) stay on `run_diffusion_benchmark.py` and may keep `server_type`.
 
 #### `mark` field
 
@@ -100,7 +101,7 @@ Recommended for L4 perf cases:
 }
 ```
 
-Multi-GPU diffusion (example: Cosmos3 with `cfg-parallel-size=2`):
+Multi-GPU generation via omni bench (example: Cosmos3 with `cfg-parallel-size=2`):
 
 ```JSON
 {
@@ -110,10 +111,17 @@ Multi-GPU diffusion (example: Cosmos3 with `cfg-parallel-size=2`):
         "full_model",
         "diffusion"
     ],
-    "server_type": "vllm-omni",
-    "benchmark_endpoint": "/v1/images/generations",
     "server_params": { "...": "..." },
-    "benchmark_params": [ { "name": "1024x1024_steps4", "...": "..." } ]
+    "benchmark_params": [
+        {
+            "name": "1024x1024_steps4",
+            "dataset_name": "random",
+            "endpoint": "/v1/images/generations",
+            "num_prompts": 3,
+            "max_concurrency": 1,
+            "extra_body": { "width": 1024, "height": 1024, "num_inference_steps": 4 }
+        }
+    ]
 }
 ```
 
@@ -132,8 +140,8 @@ Result files use the **runtime** hardware label from `get_runtime_resource_label
 
 Examples:
 
-- Omni/TTS: `result_{test_name}_{optional_hw}_{dataset}_....json` under `BENCHMARK_DIR`
-- Diffusion: one aggregate `diffusion_result_{config_stem}_{optional_hw}_{timestamp}.json` per source JSON file (array of all runs from that file)
+- Omni/TTS and OpenAI generation endpoints (`/v1/images/*`, `/v1/videos`): `result_{test_name}_{optional_hw}_{dataset}_....json` under `BENCHMARK_DIR`
+- Remaining diffusion (`run_diffusion_benchmark.py`): one aggregate `diffusion_result_{config_stem}_{optional_hw}_{timestamp}.json` per source JSON file
 
 #### Local commands
 
@@ -145,6 +153,9 @@ pytest -s -v tests/dfx/perf/scripts/run_benchmark.py -m "full_model and omni and
 # Single file (same selectors as the CI Perf steps)
 pytest -s -v tests/dfx/perf/scripts/run_diffusion_benchmark.py \
   --test-config-file tests/dfx/perf/tests/test_bagel_vllm_omni.json
+pytest -s -v tests/dfx/perf/scripts/run_benchmark.py \
+  --test-config-file tests/dfx/perf/tests/test_cosmos3_vllm_omni.json \
+  -m "H100 and B200 and cards_2"
 pytest -s -v tests/dfx/perf/scripts/run_benchmark.py \
   --test-config-file tests/dfx/perf/tests/test_qwen3_omni_async_chunk.json \
   -m "H100 and full_model and not slow"
@@ -203,7 +214,8 @@ You can add any benchmark running parameters you need here. For all optional par
 2. For boolean variables in the running parameters, modify them to forms such as ignore_eos: true/false and fill them into the JSON file.
 3. Optionally add a `baseline` object (see **Baseline thresholds** below). If you omit `baseline` or leave it empty, the performance test still runs but does not assert metric thresholds from this field.
 4. Set `"name"` on each `benchmark_params` entry for stable pytest ids and readable result keys.
-5. The qps and concurrency modes are recommended to be mutually exclusive. For detailed explanations, see the table below:
+5. Image/video generation cases (`/v1/images/generations`, `/v1/images/edits`, `/v1/videos`) use this same schema: set `endpoint` to the API path, put width/height/steps/frames in `extra_body`, use `dataset_name: random` for text-only inputs or `random-mm` when the request needs a synthetic image/video. Do not set `server_type` or `task`, and do not use `random-request-config` or kebab-case diffusion client fields.
+6. The qps and concurrency modes are recommended to be mutually exclusive. For detailed explanations, see the table below:
 
 | Parameter       | Type          | Required | Example/Values       | Description                          |
 | --------------- | ------------- | -------- | -------------------- | ------------------------------------ |
diff --git a/docs/contributing/ci/test_execution_guide.md b/docs/contributing/ci/test_execution_guide.md
index d3b02d27f73..4daf9ef8909 100644
--- a/docs/contributing/ci/test_execution_guide.md
+++ b/docs/contributing/ci/test_execution_guide.md
@@ -165,7 +165,7 @@ Failed jobs: 1/2
     pytest -sv tests/dfx/perf/scripts/run_benchmark.py -m "full_model and tts and H100"
     pytest -sv tests/dfx/perf/scripts/run_diffusion_benchmark.py -m "full_model and diffusion and H100"
     pytest -sv tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_tts.json
-    pytest -sv tests/dfx/perf/scripts/run_diffusion_benchmark.py --test-config-file tests/dfx/perf/tests/test_cosmos3_vllm_omni.json
+    pytest -sv tests/dfx/perf/scripts/run_benchmark.py --test-config-file tests/dfx/perf/tests/test_cosmos3_vllm_omni.json
     ```
     Nightly **Perf Test** jobs in [``test-nightly.yml``](https://github.com/vllm-project/vllm-omni/blob/main/.buildkite/cuda/test-nightly.yml) use ``--test-config-file`` only (no ``-m``). Weekly **Perf Test** in [``test-weekly.yml``](https://github.com/vllm-project/vllm-omni/blob/main/.buildkite/cuda/test-weekly.yml) runs ``test_qwen3_omni_vllm_text.json`` and ``test_qwen3_omni_multi_replicas.json`` (JSON ``mark`` includes ``slow``). E2e L4 function tests use ``full_model`` + ``--run-level full_model``. Example:
 
diff --git a/docs/contributing/ci/test_writing_guide.md b/docs/contributing/ci/test_writing_guide.md
index 47a62911280..da282608c45 100644
--- a/docs/contributing/ci/test_writing_guide.md
+++ b/docs/contributing/ci/test_writing_guide.md
@@ -159,7 +159,8 @@ When `mark` is present, it must be an **array** with exactly one ``hardware_mark
 }
 ```
 
-- Local bulk load: `pytest -sv tests/dfx/perf/scripts/run_diffusion_benchmark.py -m "full_model and diffusion and H100"`
+- Local bulk load: `pytest -sv tests/dfx/perf/scripts/run_benchmark.py -m "full_model and H100"` (omni/TTS and `/v1/images/*` + `/v1/videos` diffusion)
+- Diffusion chat-completions remaining cases: `pytest -sv tests/dfx/perf/scripts/run_diffusion_benchmark.py -m "full_model and diffusion and H100"`
 - Nightly CI perf steps: `--test-config-file tests/dfx/perf/tests/test_<model>_vllm_omni.json` (file selects cases; no `-m`)
 - Result filenames use **runtime** GPU detection (`get_runtime_resource_label`); `H100` is omitted on the default CI pool
 
diff --git a/recipes/LTX/LTX-2.md b/recipes/LTX/LTX-2.md
index 5a6a049eeed..b84bebd6680 100644
--- a/recipes/LTX/LTX-2.md
+++ b/recipes/LTX/LTX-2.md
@@ -5,7 +5,7 @@
 ## Pipelines
 
 | `--model-class-name` | Task | Required checkpoint repositories |
-|---|---|---|
+| --- | --- | --- |
 | `LTX2Pipeline` | LTX-2 one-stage T2V/I2V | `Lightricks/LTX-2` |
 | `LTX2TwoStagePipeline` | LTX-2 ordinary two-stage T2V/I2V | `Lightricks/LTX-2` |
 | `LTX2DistilledOneStagePipeline` | LTX-2 merged-distilled one-stage T2V/I2V | `rootonchair/LTX-2-19b-distilled` |
@@ -53,7 +53,7 @@ pipe(req, image=image, prompt=prompt)
 The consolidation also removes these registry names without aliases:
 
 | Removed name | Replacement |
-|---|---|
+| --- | --- |
 | `LTX23Pipeline` | `LTX2Pipeline`; checkpoint metadata selects LTX-2.3 |
 | `LTX2ImageToVideoPipeline` | `LTX2Pipeline` with `image=` |
 | `LTX23ImageToVideoPipeline` | `LTX2Pipeline` with `image=`; checkpoint metadata selects LTX-2.3 |
@@ -67,7 +67,7 @@ offline and serving entrypoints already use named fields and are unaffected.
 ## One-Stage Defaults
 
 | Parameter | LTX-2 | LTX-2.3 |
-|---|---:|---:|
+| --- | ---: | ---: |
 | Width × height | 768 × 512 | 768 × 512 |
 | Frames / frame rate | 121 / 24 | 121 / 24 |
 | Denoise steps | 40 | 30 |
@@ -85,7 +85,7 @@ default to `121`.
 ## Two-Stage Defaults
 
 | Parameter | Ordinary | Full-distilled |
-|---|---:|---:|
+| --- | ---: | ---: |
 | Final width × height | 1536 × 1024 | 1536 × 1024 |
 | Stage 1 width × height | 768 × 512 | 768 × 512 |
 | Frames / frame rate | 121 / 24 | 121 / 24 |
@@ -172,7 +172,7 @@ spatio-temporal guidance (STG), cross-modality guidance, and rescaling.
 Distilled stages and ordinary Stage 2 are fixed positive-only.
 
 | Parameter | Default | Effect | Alias |
-|---|---:|---|---|
+| --- | ---: | --- | --- |
 | `video_cfg_scale` | 3.0 | Video text CFG; `1.0` disables it | `video_cfg_guidance_scale` |
 | `audio_cfg_scale` | 7.0 | Audio text CFG; `1.0` disables it | `audio_cfg_guidance_scale` |
 | `video_stg_scale` | 1.0 | Video STG; `0.0` disables it | `video_stg_guidance_scale` |
@@ -210,7 +210,7 @@ per denoise step: `cond`, `uncond`, `ptb` (STG), and `mod`
 (cross-modality). The useful balanced configurations are therefore:
 
 | `--cfg-parallel-size` | Passes per rank | Guidance-slot utilization | Notes |
-|---:|---:|---:|---|
+| ---: | ---: | ---: | --- |
 | `1` | 4 | 100% | Single-rank fused guidance batch |
 | `2` | 2 | 100% | Recommended two-rank configuration |
 | `4` | 1 | 100% | One guidance pass per rank |
@@ -269,7 +269,7 @@ noted below.
 ### Complete `forward` Surface
 
 | Argument | Type/default | Meaning and constraints |
-|---|---|---|
+| --- | --- | --- |
 | `req` | `DiffusionRequestBatch`, required | Only positional argument; contains prompts and per-request sampling parameters. |
 | `image` | image or batch, `None` | Direct value wins over request images; no image selects T2V. I2V accepts one image per prompt, and a batch cannot mix T2V/I2V. |
 | `prompt` | string or list, `None` | Positive-text fallback; request prompts win. Mutually exclusive with `prompt_embeds`. |
@@ -306,7 +306,7 @@ request prompt payload; LTX guidance fields live in sampling `extra_args`.
 ### Recipe-Specific Request Capabilities
 
 | Override | One-stage | Ordinary two-stage | Distilled two-stage |
-|---|---|---|---|
+| --- | --- | --- | --- |
 | Guidance | Supported | Stage 1 only; Stage 2 is positive-only | Fixed positive-only |
 | Negative prompt/embeddings | Supported | Supported by Stage 1 | Rejected |
 | `num_inference_steps` | Supported | Controls Stage 1; Stage 2 uses 3 | Fixed at 8 for Stage 1; Stage 2 uses 3 |
@@ -357,7 +357,7 @@ bundled offline CLI do not currently expose `sigmas`.
 - The output audio sample rate comes from the loaded components and is not a
   request parameter.
 - For benchmarks, use `tests/dfx/perf/tests/test_ltx2_vllm_omni.json` with
-  `tests/dfx/perf/scripts/run_diffusion_benchmark.py`.
+  `tests/dfx/perf/scripts/run_benchmark.py`.
 - Ordinary and full-distilled two-stage T2V/I2V are supported; HQ execution
   remains out of scope.
 
diff --git a/tests/benchmarks/metrics/test_metrics.py b/tests/benchmarks/metrics/test_metrics.py
index 0aae841afa4..e7d782f9e02 100644
--- a/tests/benchmarks/metrics/test_metrics.py
+++ b/tests/benchmarks/metrics/test_metrics.py
@@ -6,11 +6,16 @@
 """
 
 import math
+from types import SimpleNamespace
 
 import pytest
 from vllm.benchmarks.serve import TaskType
 
-from vllm_omni.benchmarks.metrics.metrics import calculate_metrics
+from vllm_omni.benchmarks.metrics.metrics import (
+    aggregate_stage_durations,
+    calculate_metrics,
+    print_stage_durations_metrics,
+)
 from vllm_omni.benchmarks.patch.patch import MixRequestFuncOutput
 
 pytestmark = [pytest.mark.core_model, pytest.mark.benchmark, pytest.mark.cpu]
@@ -379,12 +384,7 @@ class _EmptyAwareTokenizer:
     """
 
     def __call__(self, text, add_special_tokens=False):
-        class _R:
-            pass
-
-        r = _R()
-        r.input_ids = [0] * len(text)
-        return r
+        return SimpleNamespace(input_ids=[0] * len(text))
 
 
 def _make_tts_output(prompt_len: int) -> MixRequestFuncOutput:
@@ -588,5 +588,80 @@ def test_image_with_generated_text_still_reports_text_result(capsys):
     assert " Image Result " in out
 
 
+def test_aggregate_stage_durations_mean_p50_p99() -> None:
+    ok_a = MixRequestFuncOutput()
+    ok_a.success = True
+    ok_a.stage_durations = {"diffuse": 1.0, "vae.decode": 0.2, "stage_0_gen_ms": 50.0}
+    ok_b = MixRequestFuncOutput()
+    ok_b.success = True
+    ok_b.stage_durations = {"diffuse": 3.0, "vae.decode": 0.4, "stage_1_gen_ms": 80.0}
+    failed = MixRequestFuncOutput()
+    failed.success = False
+    failed.stage_durations = {"diffuse": 99.0}
+
+    summaries = aggregate_stage_durations([ok_a, ok_b, failed])
+    assert summaries["stage_durations_mean"]["diffuse"] == pytest.approx(2.0)
+    assert summaries["stage_durations_p50"]["diffuse"] == pytest.approx(2.0)
+    assert summaries["stage_durations_mean"]["vae.decode"] == pytest.approx(0.3)
+    assert "stage_durations_p99" in summaries
+    assert "stage_0_gen_ms" not in summaries["stage_durations_mean"]
+    assert "stage_1_gen_ms" not in summaries["stage_durations_mean"]
+
+
+def test_print_stage_durations_metrics(capsys) -> None:
+    output = MixRequestFuncOutput()
+    output.success = True
+    output.stage_durations = {
+        "Wan22I2VPipeline.diffuse": 1.25,
+        "Wan22I2VPipeline.text_encoder.forward": 0.4,
+        "queue_wait_ms": 0.5,
+        "stage_0_gen_ms": 1000.0,
+    }
+    summaries = aggregate_stage_durations([output])
+    metrics = SimpleNamespace(**summaries)
+    print_stage_durations_metrics([99.0], metrics)
+    out = capsys.readouterr().out
+    assert "Wan22I2VPipeline" not in out
+    assert "Mean Diffuse (s):" in out
+    assert "Median Diffuse (s):" in out
+    assert "P99 Diffuse (s):" in out
+    assert "Mean Text Encoder Forward (s):" in out
+    assert "Mean Queue Wait (ms):" in out
+    assert "Stage 0 Gen" not in out
+    assert "stage_0_gen" not in out
+
+
+def test_profiler_stage_durations_print_without_print_stage(capsys) -> None:
+    output = MixRequestFuncOutput()
+    output.success = True
+    output.prompt_len = 8
+    output.latency = 1.0
+    output.stage_durations = {
+        "Wan22I2VPipeline.diffuse": 1.25,
+        "queue_wait_ms": 0.5,
+        "stage_0_gen_ms": 1000.0,
+    }
+    common = dict(
+        input_requests=[],
+        outputs=[output],
+        dur_s=1.0,
+        tokenizer=None,
+        selected_percentiles=[50.0, 99.0],
+        goodput_config_dict={},
+        task_type=TaskType.GENERATION,
+        selected_percentile_metrics=["e2el"],
+        max_concurrency=None,
+        request_rate=float("inf"),
+        benchmark_duration=1.0,
+    )
+    metrics, _ = calculate_metrics(**common, print_stage=False)
+    shown = capsys.readouterr().out
+    assert "Mean Diffuse (s):" in shown
+    assert "Mean Queue Wait (ms):" in shown
+    assert "Stage 0 Gen" not in shown
+    assert metrics.stage_durations_mean["Wan22I2VPipeline.diffuse"] == pytest.approx(1.25)
+    assert "stage_0_gen_ms" not in metrics.stage_durations_mean
+
+
 if __name__ == "__main__":
     pytest.main([__file__, "-v", "-s"])
diff --git a/tests/benchmarks/patch/test_patch.py b/tests/benchmarks/patch/test_patch.py
index d2552bbff53..b203fce59d2 100644
--- a/tests/benchmarks/patch/test_patch.py
+++ b/tests/benchmarks/patch/test_patch.py
@@ -22,12 +22,17 @@
 )
 from vllm_omni.benchmarks.patch.patch import (
     MixRequestFuncOutput,
+    _add_combined_video_form_references,
     _add_video_extra_body_to_form,
     _add_video_reference_to_form,
+    _apply_image_metrics_from_payload,
     _apply_stage0_token_timings,
     _apply_video_metrics_from_payload,
     _attach_seed_tts_to_request_func_input,
     _build_benchmark_session,
+    _extract_stage_durations_from_payload,
+    _iter_image_reference_inputs,
+    _iter_video_reference_inputs,
     _omni_request_timeout_s,
     async_request_openai_chat_omni_completions,
     async_request_openai_image_edits_omni,
@@ -1359,16 +1364,15 @@ async def __aexit__(self, *_args):
             return None
 
     captured_stream: list[str] = []
-    real_add_field = None
+    import aiohttp
+
+    real_add_field = aiohttp.FormData.add_field
 
     def tracking_add_field(self, name, value=None, **kwargs):
         if name == "stream":
             captured_stream.append(str(value))
         return real_add_field(self, name, value, **kwargs)
 
-    import aiohttp
-
-    real_add_field = aiohttp.FormData.add_field
     mocker.patch.object(aiohttp.FormData, "add_field", tracking_add_field)
 
     mock_session = mocker.AsyncMock()
@@ -1512,8 +1516,7 @@ def tracking_add_field(self, name, value=None, **kwargs):
     "reference",
     [
         {"image_url": "https://example.com/ref.png"},
-        {"file_id": "file-abc"},
-        [{"image_url": "https://example.com/a.png"}, {"file_id": "file-xyz"}],
+        [{"image_url": "https://example.com/a.png"}, {"image_url": "https://example.com/b.png"}],
     ],
 )
 def test_video_structured_image_reference_serialized_to_form(reference: object, mocker: MockerFixture) -> None:
@@ -1541,18 +1544,395 @@ def tracking_add_field(self, name, value=None, **kwargs):
     assert field_names.count("image_reference") == 1
     assert "input_reference" not in field_names
     payload = next(value for name, value in captured if name == "image_reference")
+    assert isinstance(payload, (str, bytes, bytearray))
     assert json.loads(payload) == reference
 
 
+def test_video_file_id_reference_is_rejected() -> None:
+    """file_id is unsupported on the server and must not be sent as image_reference."""
+    import aiohttp
+
+    form = aiohttp.FormData()
+    with pytest.raises(ValueError, match="file_id is not supported yet"):
+        _add_video_reference_to_form(form, {"file_id": "file-abc"})
+    with pytest.raises(ValueError, match="file_id is not supported yet"):
+        _add_video_reference_to_form(form, [{"image_url": "https://example.com/a.png"}, {"file_id": "file-xyz"}])
+    with pytest.raises(ValueError, match="file_id is not supported yet"):
+        _add_combined_video_form_references(form, None, {"video_reference": {"file_id": "file-vid"}})
+
+
+def test_image_reference_urls_from_random_mm_content(mocker: MockerFixture) -> None:
+    """random-mm image_url parts keep an explicit image_reference type."""
+    import aiohttp
+
+    content = [
+        {
+            "type": "image_url",
+            "image_url": {"url": "https://example.com/ref.png"},
+        },
+        {
+            "type": "image_url",
+            "image_url": {"url": f"data:image/png;base64,{_MIN_PNG_B64}"},
+        },
+    ]
+    refs = list(_iter_image_reference_inputs(content))
+    assert refs == [
+        {"image_url": "https://example.com/ref.png"},
+        {"image_url": f"data:image/png;base64,{_MIN_PNG_B64}"},
+    ]
+
+    captured: list[tuple[str, object]] = []
+    real_add_field = aiohttp.FormData.add_field
+
+    def tracking_add_field(self, name, value=None, **kwargs):
+        captured.append((str(name), value))
+        return real_add_field(self, name, value, **kwargs)
+
+    mocker.patch.object(aiohttp.FormData, "add_field", tracking_add_field)
+    form = aiohttp.FormData()
+    assert _add_video_reference_to_form(form, refs[0]) is True
+    assert _add_video_reference_to_form(form, refs[1]) is True
+
+    field_names = [name for name, _ in captured]
+    assert field_names.count("image_reference") == 2
+    assert "video_reference" not in field_names
+    payloads = []
+    for name, value in captured:
+        if name != "image_reference":
+            continue
+        assert isinstance(value, (str, bytes, bytearray))
+        payloads.append(json.loads(value))
+    assert payloads == refs
+
+
+def test_video_reference_urls_from_random_mm_content(mocker: MockerFixture) -> None:
+    """random-mm data:video_url parts stay on video_reference."""
+    import aiohttp
+
+    content = [
+        {
+            "type": "video_url",
+            "video_url": {"url": "data:video/mp4;base64,AAAA"},
+        }
+    ]
+    refs = list(_iter_video_reference_inputs(content))
+    assert refs == [{"video_url": "data:video/mp4;base64,AAAA"}]
+
+    captured: list[tuple[str, object]] = []
+    real_add_field = aiohttp.FormData.add_field
+
+    def tracking_add_field(self, name, value=None, **kwargs):
+        captured.append((str(name), value))
+        return real_add_field(self, name, value, **kwargs)
+
+    mocker.patch.object(aiohttp.FormData, "add_field", tracking_add_field)
+    form = aiohttp.FormData()
+    assert _add_video_reference_to_form(form, refs[0]) is True
+
+    field_names = [name for name, _ in captured]
+    assert field_names.count("video_reference") == 1
+    assert "input_references" not in field_names
+    assert "input_reference" not in field_names
+    assert "image_reference" not in field_names
+    payload = next(value for name, value in captured if name == "video_reference")
+    assert isinstance(payload, (str, bytes, bytearray))
+    assert json.loads(payload) == {"video_url": "data:video/mp4;base64,AAAA"}
+
+
+def test_video_reference_https_url_from_random_mm_content(mocker: MockerFixture) -> None:
+    """HTTP(S) video_url parts must stay on video_reference, not image_reference."""
+    import aiohttp
+
+    content = [
+        {
+            "type": "video_url",
+            "video_url": {"url": "https://example.com/ref.mp4"},
+        }
+    ]
+    refs = list(_iter_video_reference_inputs(content))
+    assert refs == [{"video_url": "https://example.com/ref.mp4"}]
+
+    captured: list[tuple[str, object]] = []
+    real_add_field = aiohttp.FormData.add_field
+
+    def tracking_add_field(self, name, value=None, **kwargs):
+        captured.append((str(name), value))
+        return real_add_field(self, name, value, **kwargs)
+
+    mocker.patch.object(aiohttp.FormData, "add_field", tracking_add_field)
+    form = aiohttp.FormData()
+    assert _add_video_reference_to_form(form, refs[0]) is True
+
+    field_names = [name for name, _ in captured]
+    assert field_names.count("video_reference") == 1
+    assert "image_reference" not in field_names
+    assert "input_references" not in field_names
+    payload = next(value for name, value in captured if name == "video_reference")
+    assert isinstance(payload, (str, bytes, bytearray))
+    assert json.loads(payload) == {"video_url": "https://example.com/ref.mp4"}
+
+
+@pytest.mark.parametrize(
+    ("reference", "field_name", "expected"),
+    [
+        ("data:image/png;base64,AAAA", "image_reference", {"image_url": "data:image/png;base64,AAAA"}),
+        ("data:video/mp4;base64,AAAA", "video_reference", {"video_url": "data:video/mp4;base64,AAAA"}),
+        ("https://example.com/ref.png", "image_reference", {"image_url": "https://example.com/ref.png"}),
+        ("https://example.com/ref.mp4", "video_reference", {"video_url": "https://example.com/ref.mp4"}),
+    ],
+)
+def test_bare_reference_string_uses_matching_json_field(
+    reference: str,
+    field_name: str,
+    expected: dict[str, str],
+    mocker: MockerFixture,
+) -> None:
+    """Image and video URLs use matching JSON fields; neither is a file upload."""
+    import aiohttp
+
+    captured: list[tuple[str, object]] = []
+    real_add_field = aiohttp.FormData.add_field
+
+    def tracking_add_field(self, name, value=None, **kwargs):
+        captured.append((str(name), value))
+        return real_add_field(self, name, value, **kwargs)
+
+    mocker.patch.object(aiohttp.FormData, "add_field", tracking_add_field)
+    assert _add_video_reference_to_form(aiohttp.FormData(), reference) is True
+    assert [name for name, _ in captured] == [field_name]
+    payload = captured[0][1]
+    assert isinstance(payload, (str, bytes, bytearray))
+    assert json.loads(payload) == expected
+
+
+def test_bare_video_string_is_not_collected_as_image() -> None:
+    content = ["https://example.com/ref.mp4", "data:video/mp4;base64,AAAA"]
+    assert list(_iter_image_reference_inputs(content)) == []
+    assert list(_iter_video_reference_inputs(content)) == [
+        {"video_url": "https://example.com/ref.mp4"},
+        {"video_url": "data:video/mp4;base64,AAAA"},
+    ]
+
+
+def test_bare_http_without_media_extension_is_rejected() -> None:
+    import aiohttp
+
+    with pytest.raises(ValueError, match="image or video extension"):
+        _add_video_reference_to_form(aiohttp.FormData(), "https://example.com/ref")
+
+
+def test_image_and_inline_video_use_combined_reference_fields(mocker: MockerFixture) -> None:
+    """Image plus data:video must be image_reference + video_reference, not input_references."""
+    import aiohttp
+
+    content = [
+        {"type": "image_url", "image_url": {"url": "https://example.com/ref.png"}},
+        {"type": "video_url", "video_url": {"url": "data:video/mp4;base64,AAAA"}},
+    ]
+    captured: list[tuple[str, object]] = []
+    real_add_field = aiohttp.FormData.add_field
+
+    def tracking_add_field(self, name, value=None, **kwargs):
+        captured.append((str(name), value))
+        return real_add_field(self, name, value, **kwargs)
+
+    mocker.patch.object(aiohttp.FormData, "add_field", tracking_add_field)
+    form = aiohttp.FormData()
+    _add_combined_video_form_references(form, content)
+
+    field_names = [name for name, _ in captured]
+    assert field_names.count("image_reference") == 1
+    assert field_names.count("video_reference") == 1
+    assert "input_references" not in field_names
+    assert "input_reference" not in field_names
+    image_payload = next(value for name, value in captured if name == "image_reference")
+    video_payload = next(value for name, value in captured if name == "video_reference")
+    assert isinstance(image_payload, (str, bytes, bytearray))
+    assert isinstance(video_payload, (str, bytes, bytearray))
+    assert json.loads(image_payload) == {"image_url": "https://example.com/ref.png"}
+    assert json.loads(video_payload) == {"video_url": "data:video/mp4;base64,AAAA"}
+
+
+def test_image_url_and_bare_data_video_string_use_json_fields(mocker: MockerFixture) -> None:
+    """Image URL plus a bare data:video string must not upload input_references."""
+    import aiohttp
+
+    content = [{"type": "image_url", "image_url": {"url": "https://example.com/ref.png"}}]
+    extra_body = {"video_reference": "data:video/mp4;base64,AAAA"}
+    captured: list[tuple[str, object]] = []
+    real_add_field = aiohttp.FormData.add_field
+
+    def tracking_add_field(self, name, value=None, **kwargs):
+        captured.append((str(name), value))
+        return real_add_field(self, name, value, **kwargs)
+
+    mocker.patch.object(aiohttp.FormData, "add_field", tracking_add_field)
+    form = aiohttp.FormData()
+    _add_combined_video_form_references(form, content, extra_body)
+
+    field_names = [name for name, _ in captured]
+    assert field_names.count("image_reference") == 1
+    assert field_names.count("video_reference") == 1
+    assert "input_references" not in field_names
+    assert "input_reference" not in field_names
+    image_payload = next(value for name, value in captured if name == "image_reference")
+    video_payload = next(value for name, value in captured if name == "video_reference")
+    assert isinstance(image_payload, (str, bytes, bytearray))
+    assert isinstance(video_payload, (str, bytes, bytearray))
+    assert json.loads(image_payload) == {"image_url": "https://example.com/ref.png"}
+    assert json.loads(video_payload) == {"video_url": "data:video/mp4;base64,AAAA"}
+
+    captured.clear()
+    assert _add_video_reference_to_form(form, "data:video/mp4;base64,AAAA") is True
+    assert [name for name, _ in captured] == ["video_reference"]
+
+
+def _oversized_data_video_url() -> str:
+    """data:video whose JSON text part is larger than 1MB."""
+    return "data:video/mp4;base64," + ("A" * (1024 * 1024))
+
+
+def test_oversized_data_video_uploads_as_input_references(mocker: MockerFixture) -> None:
+    """A lone data:video over the 1MB text limit is uploaded, not sent as JSON."""
+    import aiohttp
+
+    video_url = _oversized_data_video_url()
+    captured: list[tuple[str, object]] = []
+    real_add_field = aiohttp.FormData.add_field
+
+    def tracking_add_field(self, name, value=None, **kwargs):
+        captured.append((str(name), value))
+        return real_add_field(self, name, value, **kwargs)
+
+    mocker.patch.object(aiohttp.FormData, "add_field", tracking_add_field)
+    assert _add_video_reference_to_form(aiohttp.FormData(), {"video_url": video_url}) is True
+    uploaded = next(value for name, value in captured if name == "input_references")
+    assert uploaded == base64.b64decode("A" * (1024 * 1024))
+    field_names = [name for name, _ in captured]
+    assert "video_reference" not in field_names
+    assert "image_reference" not in field_names
+
+
+def test_oversized_data_video_with_image_stays_json(mocker: MockerFixture) -> None:
+    """input_references cannot be combined with image_reference, even for a large video."""
+    import aiohttp
+
+    content = [{"type": "image_url", "image_url": {"url": "https://example.com/ref.png"}}]
+    extra_body = {"video_reference": _oversized_data_video_url()}
+    captured: list[tuple[str, object]] = []
+    real_add_field = aiohttp.FormData.add_field
+
+    def tracking_add_field(self, name, value=None, **kwargs):
+        captured.append((str(name), value))
+        return real_add_field(self, name, value, **kwargs)
+
+    mocker.patch.object(aiohttp.FormData, "add_field", tracking_add_field)
+    _add_combined_video_form_references(aiohttp.FormData(), content, extra_body)
+    field_names = [name for name, _ in captured]
+    assert "image_reference" in field_names
+    assert "video_reference" in field_names
+    assert "input_references" not in field_names
+
+
+def test_image_upload_bytes_and_structured_video_use_json_fields(mocker: MockerFixture) -> None:
+    """Image upload bytes plus a video URL must not use singular input_reference."""
+    import aiohttp
+
+    image_bytes = base64.b64decode(_MIN_PNG_B64)
+    extra_body = {
+        "image_reference": {"bytes": image_bytes, "content_type": "image/png"},
+        "video_reference": {"video_url": "https://example.com/ref.mp4"},
+    }
+    captured: list[tuple[str, object]] = []
+    real_add_field = aiohttp.FormData.add_field
+
+    def tracking_add_field(self, name, value=None, **kwargs):
+        captured.append((str(name), value))
+        return real_add_field(self, name, value, **kwargs)
+
+    mocker.patch.object(aiohttp.FormData, "add_field", tracking_add_field)
+    form = aiohttp.FormData()
+    _add_combined_video_form_references(form, None, extra_body)
+
+    field_names = [name for name, _ in captured]
+    assert field_names.count("image_reference") == 1
+    assert field_names.count("video_reference") == 1
+    assert "input_reference" not in field_names
+    assert "input_references" not in field_names
+    image_payload = next(value for name, value in captured if name == "image_reference")
+    video_payload = next(value for name, value in captured if name == "video_reference")
+    assert isinstance(image_payload, (str, bytes, bytearray))
+    assert isinstance(video_payload, (str, bytes, bytearray))
+    assert json.loads(image_payload) == {"image_url": f"data:image/png;base64,{_MIN_PNG_B64}"}
+    assert json.loads(video_payload) == {"video_url": "https://example.com/ref.mp4"}
+
+
 def test_video_unsupported_image_reference_raises() -> None:
     import aiohttp
 
     form = aiohttp.FormData()
-    with pytest.raises(ValueError, match="Unsupported image_reference"):
+    with pytest.raises(ValueError, match="Unsupported reference"):
         _add_video_reference_to_form(form, {"not_a_supported_key": "x"})
-    with pytest.raises(ValueError, match="Unsupported image_reference"):
+    with pytest.raises(ValueError, match="Unsupported reference"):
         _add_video_reference_to_form(form, "/tmp/does-not-exist-ref.png")
 
 
+def test_extract_stage_durations_from_video_and_image_shapes() -> None:
+    video_payload = {
+        "stage_durations": {
+            "diffuse": 1.5,
+            "text_encoder.forward": 0.2,
+            "vae.decode": 0.1,
+            "stage_0_gen_ms": 4000.0,
+        }
+    }
+    assert _extract_stage_durations_from_payload(video_payload) == video_payload["stage_durations"]
+
+    image_stage_durations = {
+        "diffuse": 2.0,
+        "vae.decode": 0.3,
+    }
+    image_payload = {
+        "metrics": {"stage_durations": image_stage_durations},
+        "data": [{"b64_json": "x"}],
+    }
+    assert _extract_stage_durations_from_payload(image_payload) == image_stage_durations
+
+
+def test_video_metrics_persist_full_stage_durations() -> None:
+    output = MixRequestFuncOutput()
+    output.latency = 4.2
+    stage_durations = {
+        "diffuse": 1.5,
+        "vae.decode": 0.25,
+        "stage_0_gen_ms": 4000.0,
+    }
+    payload = {
+        "duration_s": 2.0,
+        "num_frames": 48,
+        "fps": 24.0,
+        "stage_durations": stage_durations,
+    }
+    _apply_video_metrics_from_payload(output, payload, {})
+    assert output.stage_durations == stage_durations
+    assert output.video_generation_time_ms == pytest.approx(4000.0)
+
+
+def test_image_metrics_persist_stage_durations_from_metrics() -> None:
+    output = MixRequestFuncOutput()
+    stage_durations = {
+        "diffuse": 1.1,
+        "text_encoder.forward": 0.4,
+        "vae.decode": 0.2,
+    }
+    payload = {
+        "created": 1,
+        "data": [{"b64_json": _MIN_PNG_B64}],
+        "metrics": {"stage_durations": stage_durations},
+    }
+    assert _apply_image_metrics_from_payload(output, payload) == 1
+    assert output.stage_durations == stage_durations
+
+
 if __name__ == "__main__":
     pytest.main([__file__, "-v", "-s"])
diff --git a/tests/dfx/conftest.py b/tests/dfx/conftest.py
index 864fb7ab382..2015f6389b7 100644
--- a/tests/dfx/conftest.py
+++ b/tests/dfx/conftest.py
@@ -1,3 +1,6 @@
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM-Omni project
+
 import json
 import os
 import re
@@ -84,17 +87,20 @@ def resolve_pytest_marks(mark_field: Any) -> list[pytest.MarkDecorator]:
     raise ValueError(f"mark must be a list; got {type(mark_field).__name__}")
 
 
-def _mark_names(mark_field: Any) -> set[str]:
-    if isinstance(mark_field, list):
-        return {str(item) for item in mark_field if isinstance(item, str)}
-    return set()
+def is_diffusion_perf_config(cfg: dict[str, Any]) -> bool:
+    """True for perf JSON cases intended for ``run_diffusion_benchmark.py``.
 
+    Schema split (not marks / endpoints):
 
-def is_diffusion_perf_config(cfg: dict[str, Any]) -> bool:
-    """True for perf JSON cases intended for ``run_diffusion_benchmark.py``."""
-    if cfg.get("server_type") is not None:
-        return True
-    return "diffusion" in _mark_names(cfg.get("mark"))
+    - ``benchmark_params[].dataset`` → diffusion client (``run_diffusion_benchmark.py``)
+    - ``benchmark_params[].dataset_name`` → ``vllm bench serve --omni`` (``run_benchmark.py``)
+    """
+    for params in cfg.get("benchmark_params") or []:
+        if not isinstance(params, dict):
+            continue
+        if "dataset" in params and "dataset_name" not in params:
+            return True
+    return False
 
 
 def _marks_by_test_name(configs: list[dict[str, Any]]) -> dict[str, list[pytest.MarkDecorator]]:
@@ -654,6 +660,7 @@ def run_benchmark(
     random_output_len: Any | None = None,
     resource_label: str | None = None,
     num_warmups: int = 2,
+    benchmark_params_name: str | None = None,
 ) -> dict[str, Any]:
     """Run one ``vllm bench serve --omni`` iteration and return parsed metrics.
 
@@ -662,15 +669,35 @@ def run_benchmark(
     maps keep every hardware bucket, but each metric is reduced to the value for
     this concurrency / request-rate step. If the benchmark exits without writing a
     result file, ``result_omni_template.json`` is used as a fallback.
+
+    ``benchmark_params_name`` (the ``name`` field from ``benchmark_params``) is
+    persisted in the result JSON and filename so multiple generation configs
+    under one ``test_name`` stay distinguishable (e.g. Wan USP2 832x480 vs
+    1280x720).
     """
     current_dt = datetime.now().strftime("%Y%m%d-%H%M%S")
     ri = _safe_filename_token(random_input_len)
     ro = _safe_filename_token(random_output_len)
     hw = resource_label_for_filename(resource_label)
+    name_token = _safe_filename_token(benchmark_params_name) if benchmark_params_name else ""
+    if name_token == "na":
+        name_token = ""
+    name_parts = [f"result_{test_name}"]
     if hw:
-        result_filename = f"result_{test_name}_{hw}_{dataset_name}_{flow}_{num_prompt}_in{ri}_out{ro}_{current_dt}.json"
-    else:
-        result_filename = f"result_{test_name}_{dataset_name}_{flow}_{num_prompt}_in{ri}_out{ro}_{current_dt}.json"
+        name_parts.append(hw)
+    if name_token:
+        name_parts.append(name_token)
+    name_parts.extend(
+        [
+            str(dataset_name),
+            str(flow),
+            str(num_prompt),
+            f"in{ri}",
+            f"out{ro}",
+            current_dt,
+        ]
+    )
+    result_filename = "_".join(name_parts) + ".json"
     if "--result-filename" in args:
         print(f"The result file will be overwritten by {result_filename}")
     command = (
@@ -740,6 +767,9 @@ def _forward_stream(stream) -> None:
     if random_output_len is not None:
         result["random_output_len"] = random_output_len
     result["Hardware"] = hardware_json_value(resource_label)
+    result["test_name"] = test_name
+    if benchmark_params_name:
+        result["name"] = benchmark_params_name
     with open(result_path, "w", encoding="utf-8") as f:
         json.dump(result, f, ensure_ascii=False, indent=2)
     return result
diff --git a/tests/dfx/perf/scripts/run_benchmark.py b/tests/dfx/perf/scripts/run_benchmark.py
index 94df2edc55e..5648ffff5f2 100644
--- a/tests/dfx/perf/scripts/run_benchmark.py
+++ b/tests/dfx/perf/scripts/run_benchmark.py
@@ -50,12 +50,19 @@ def _get_config_file_from_argv() -> str | None:
     _all_configs = load_benchmark_configs(config_dir=_PERF_TESTS_DIR)
     BENCHMARK_CONFIGS = [cfg for cfg in _all_configs if not is_diffusion_perf_config(cfg)]
     print(
-        f"No --test-config-file: loaded {len(BENCHMARK_CONFIGS)} omni/tts case(s) from "
+        f"No --test-config-file: loaded {len(BENCHMARK_CONFIGS)} omni/tts/generation case(s) from "
         f"{_PERF_TESTS_DIR}/*.json (skipped {len(_all_configs) - len(BENCHMARK_CONFIGS)} diffusion; "
         f"use -m to filter, e.g. -m tts)"
     )
 else:
-    BENCHMARK_CONFIGS = load_benchmark_configs(CONFIG_FILE_PATH)
+    _loaded = load_benchmark_configs(CONFIG_FILE_PATH)
+    BENCHMARK_CONFIGS = [cfg for cfg in _loaded if not is_diffusion_perf_config(cfg)]
+    skipped = len(_loaded) - len(BENCHMARK_CONFIGS)
+    if skipped:
+        print(
+            f"--test-config-file: loaded {len(BENCHMARK_CONFIGS)} omni/tts/generation case(s); "
+            f"skipped {skipped} remaining diffusion case(s) (chat completions / custom jsonl)"
+        )
 
 DEPLOY_CONFIGS_DIR = Path(__file__).parent.parent / "deploy"
 server_to_benchmark_mapping = create_test_parameter_mapping(BENCHMARK_CONFIGS)
@@ -93,24 +100,107 @@ def close(self) -> None:
             stack.close()
 
 
+# OmniServer defaults for flags not already present in JSON ``serve_args`` /
+# ``extra_cli_args``. Add new (flag, value) pairs here rather than special-casing.
+_OMNI_DEFAULT_SERVER_ARGS: tuple[tuple[str, str], ...] = (
+    ("--stage-init-timeout", "600"),
+    ("--init-timeout", "900"),
+)
+
+
+def _cli_flag_names(cli_args: tuple[str, ...] | list[str]) -> set[str]:
+    """Return long-option names present in a flat CLI argv list."""
+    names: set[str] = set()
+    for item in cli_args:
+        token = str(item)
+        if not token.startswith("--"):
+            continue
+        names.add(token.split("=", 1)[0])
+    return names
+
+
+def _merge_omni_default_server_args(
+    extra_cli_args: tuple[str, ...] | list[str],
+    *,
+    use_omni: bool,
+    defaults: tuple[tuple[str, str], ...] = _OMNI_DEFAULT_SERVER_ARGS,
+) -> list[str]:
+    """Fill Omni defaults for flags not already set in JSON-derived CLI args.
+
+    JSON ``serve_args`` / ``extra_cli_args`` win; only missing flags are appended.
+    """
+    if not use_omni:
+        return []
+    present = _cli_flag_names(extra_cli_args)
+    args: list[str] = []
+    for flag, value in defaults:
+        if flag not in present:
+            args += [flag, value]
+    return args
+
+
+def _omni_server_env() -> dict[str, str]:
+    """Writable video/image storage for ``/v1/videos`` and related generation APIs."""
+    result_dir = Path(os.environ.get("BENCHMARK_DIR", "tests/dfx/perf/results"))
+    storage_path = Path(os.environ.get("VLLM_OMNI_STORAGE_PATH", str(result_dir / "storage")))
+    storage_path.mkdir(parents=True, exist_ok=True)
+    return {"VLLM_OMNI_STORAGE_PATH": str(storage_path)}
+
+
+def _resolve_offline_model(model: str) -> str:
+    """Resolve HF ids / MiniMax env overrides the same way as the diffusion runner."""
+    import huggingface_hub
+
+    from vllm_omni.transformers_utils.repo_utils import hf_api
+
+    if not model or os.path.isdir(model):
+        return model
+
+    model_env_overrides = {
+        "MiniMaxAI/MiniMax-H3": "VLLM_TEST_MINIMAX_H3_MODEL",
+        "MiniMaxAI/MiniMax-H3/FL2VA": "VLLM_TEST_MINIMAX_H3_FL2VA_MODEL",
+        "MiniMaxAI/MiniMax-H3/Ref2VA": "VLLM_TEST_MINIMAX_H3_REF2VA_MODEL",
+    }
+    env_name = model_env_overrides.get(model)
+    if env_name:
+        env_model = os.environ.get(env_name)
+        if env_model:
+            return env_model
+
+    parts = model.split("/")
+    if len(parts) >= 3:
+        repo_id = "/".join(parts[:2])
+        subfolder = "/".join(parts[2:])
+        snapshot_root = hf_api().snapshot_download(
+            repo_id,
+            allow_patterns=[f"{subfolder}/**"],
+            local_files_only=huggingface_hub.constants.HF_HUB_OFFLINE,
+        )
+        return str(Path(snapshot_root) / subfolder)
+
+    if not huggingface_hub.constants.HF_HUB_OFFLINE:
+        return model
+    return hf_api().snapshot_download(model, local_files_only=True)
+
+
 @contextmanager
 def _start_omni_server(server_param):
     test_name, model, stage_config_path, stage_overrides, extra_cli_args, use_omni = server_param
+    extra = tuple(extra_cli_args or ())
+    model = _resolve_offline_model(model)
 
     print(f"Starting OmniServer with test: {test_name}, model: {model}")
 
-    server_args: list[str] = []
-    if use_omni:
-        server_args += ["--stage-init-timeout", "600", "--init-timeout", "900"]
+    server_args: list[str] = _merge_omni_default_server_args(extra, use_omni=use_omni)
     # --deploy-config and --stage-overrides compose at the CLI (see vllm_omni/entrypoints/utils.py):
     # deploy-config sets the base; stage-overrides are applied on top. Both can be set.
     if stage_config_path:
         server_args = ["--deploy-config", stage_config_path] + server_args
     if stage_overrides:
         server_args = ["--stage-overrides", stage_overrides] + server_args
-    if extra_cli_args:
-        server_args = list(extra_cli_args) + server_args
-    with OmniServer(model, server_args, use_omni=use_omni) as server:
+    if extra:
+        server_args = list(extra) + server_args
+    with OmniServer(model, server_args, use_omni=use_omni, env_dict=_omni_server_env()) as server:
         server.test_name = test_name
         print("OmniServer started successfully")
         yield server
@@ -259,11 +349,24 @@ def to_list(value, default=None):
         "eval_phase",
         "trust_remote_code",
         "expected_duplex_audio_turns_per_session",
+        "name",
+        "enable_negative_prompt",
+        "random_request_config",
+        "num_input_images",
+        "warmup_requests",
+        "warmup_concurrency",
+        "warmup_num_inference_steps",
     }
 
+    param_keys = {str(key).replace("-", "_") for key in params}
+    if "model" not in param_keys:
+        args.extend(["--model", str(model)])
+
     for key, value in params.items():
         if key in exclude_keys or value is None:
             continue
+        if key in {"extra_body", "extra-body"} and value == {}:
+            continue
 
         arg_name = f"--{key.replace('_', '-')}"
 
@@ -298,6 +401,7 @@ def to_list(value, default=None):
             random_output_len=params.get("random_output_len"),
             resource_label=resource_label,
             num_warmups=_resolve_num_warmups(params, default=2),
+            benchmark_params_name=params.get("name") if isinstance(params.get("name"), str) else None,
         )
         assert_result(result, params, num_prompt)
 
@@ -316,5 +420,6 @@ def to_list(value, default=None):
             random_output_len=params.get("random_output_len"),
             resource_label=resource_label,
             num_warmups=_resolve_num_warmups(params, default=max(2, int(concurrency))),
+            benchmark_params_name=params.get("name") if isinstance(params.get("name"), str) else None,
         )
         assert_result(result, params, num_prompt)
diff --git a/tests/dfx/perf/scripts/run_diffusion_benchmark.py b/tests/dfx/perf/scripts/run_diffusion_benchmark.py
index ffcfc45b721..0523b9b3a1a 100644
--- a/tests/dfx/perf/scripts/run_diffusion_benchmark.py
+++ b/tests/dfx/perf/scripts/run_diffusion_benchmark.py
@@ -204,11 +204,11 @@ def load_diffusion_benchmark_configs(
 ) -> list[dict[str, Any]]:
     """Load one diffusion benchmark JSON, or merge all ``*.json`` under *config_dir*."""
     if config_path is not None:
-        configs = load_configs(config_path)
+        loaded = load_configs(config_path)
         source = str(Path(config_path).resolve())
-        for cfg in configs:
+        for cfg in loaded:
             cfg.setdefault(_DIFFUSION_SOURCE_CONFIG_KEY, source)
-        return configs
+        return loaded
     if config_dir is None:
         raise ValueError("load_diffusion_benchmark_configs requires config_path or config_dir")
     configs: list[dict[str, Any]] = []
@@ -231,7 +231,15 @@ def load_diffusion_benchmark_configs(
         f"use -m to filter, e.g. -m diffusion)"
     )
 else:
-    BENCHMARK_CONFIGS = load_diffusion_benchmark_configs(CONFIG_FILE_PATH)
+    _loaded = load_diffusion_benchmark_configs(CONFIG_FILE_PATH)
+    BENCHMARK_CONFIGS = [cfg for cfg in _loaded if is_diffusion_perf_config(cfg)]
+    skipped = len(_loaded) - len(BENCHMARK_CONFIGS)
+    if skipped:
+        print(
+            f"--test-config-file: loaded {len(BENCHMARK_CONFIGS)} diffusion case(s); "
+            f"skipped {skipped} omni-bench generation case(s) "
+            f"(/v1/images/edits, /v1/images/generations, /v1/videos → run_benchmark.py)"
+        )
 
 _AGGREGATED_RESULT_FILES_BY_SOURCE: dict[str, Path] = {}
 
diff --git a/tests/dfx/perf/tests/test_cosmos3_vllm_omni.json b/tests/dfx/perf/tests/test_cosmos3_vllm_omni.json
index 5e2b9d3dc7f..662f845f648 100644
--- a/tests/dfx/perf/tests/test_cosmos3_vllm_omni.json
+++ b/tests/dfx/perf/tests/test_cosmos3_vllm_omni.json
@@ -5,7 +5,10 @@
       {
         "hardware_marks": {
           "res": {
-            "cuda": ["H100", "B200"]
+            "cuda": [
+              "H100",
+              "B200"
+            ]
           },
           "num_cards": 2
         }
@@ -14,8 +17,6 @@
       "diffusion"
     ],
     "description": "Cosmos3-Nano text-to-image official demo workload on 2 GPUs",
-    "server_type": "vllm-omni",
-    "benchmark_endpoint": "/v1/images/generations",
     "server_params": {
       "model": "nvidia/Cosmos3-Nano",
       "serve_args": {
@@ -34,24 +35,28 @@
     "benchmark_params": [
       {
         "name": "1024x1024_steps4",
-        "dataset": "random",
-        "task": "t2i",
-        "width": 1024,
-        "height": 1024,
-        "num-inference-steps": 4,
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "seed": 42,
-        "extra-body": {
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/images/generations",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 1024,
+          "height": 1024,
+          "num_inference_steps": 4,
+          "seed": 42,
           "guidance_scale": 7.0,
           "flow_shift": 3.0,
           "guardrails": false
         },
         "baseline": {
           "H100": {
-            "throughput_qps": 1.1958,
-            "latency_mean": 0.843,
-            "peak_memory_mb_mean": 80000.0
+            "request_throughput": 1.1958,
+            "mean_e2el_ms": 843.0,
+            "mean_peak_memory_mb": 80000.0
           }
         }
       }
@@ -63,7 +68,10 @@
       {
         "hardware_marks": {
           "res": {
-            "cuda": ["H100", "B200"]
+            "cuda": [
+              "H100",
+              "B200"
+            ]
           },
           "num_cards": 2
         }
@@ -72,8 +80,6 @@
       "diffusion"
     ],
     "description": "Cosmos3-Nano text-to-video official demo workload on 2 GPUs",
-    "server_type": "vllm-omni",
-    "benchmark_endpoint": "/v1/videos",
     "server_params": {
       "model": "nvidia/Cosmos3-Nano",
       "serve_args": {
@@ -92,23 +98,21 @@
     "benchmark_params": [
       {
         "name": "1280x720_frames189_steps4",
-        "dataset": "random",
-        "task": "t2v",
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "seed": 42,
-        "random-request-config": [
-          {
-            "width": 1280,
-            "height": 720,
-            "prompt": "A robot arm is cleaning a plate in the kitchen.",
-            "num_inference_steps": 4,
-            "num_frames": 189,
-            "fps": 24,
-            "weight": 1
-          }
-        ],
-        "extra-body": {
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 1280,
+          "height": 720,
+          "num_inference_steps": 4,
+          "num_frames": 189,
+          "fps": 24,
+          "seed": 42,
           "guidance_scale": 6.0,
           "flow_shift": 10.0,
           "max_sequence_length": 4096,
@@ -118,9 +122,9 @@
         },
         "baseline": {
           "H100": {
-            "throughput_qps": 0.0307,
-            "latency_mean": 32.7665,
-            "peak_memory_mb_mean": 28762.4762
+            "request_throughput": 0.0307,
+            "mean_e2el_ms": 32766.5,
+            "mean_peak_memory_mb": 28762.4762
           }
         }
       }
@@ -132,7 +136,10 @@
       {
         "hardware_marks": {
           "res": {
-            "cuda": ["H100", "B200"]
+            "cuda": [
+              "H100",
+              "B200"
+            ]
           },
           "num_cards": 2
         }
@@ -141,8 +148,6 @@
       "diffusion"
     ],
     "description": "Cosmos3-Nano image-to-video official demo workload on 2 GPUs",
-    "server_type": "vllm-omni",
-    "benchmark_endpoint": "/v1/videos",
     "server_params": {
       "model": "nvidia/Cosmos3-Nano",
       "serve_args": {
@@ -161,24 +166,30 @@
     "benchmark_params": [
       {
         "name": "1280x720_frames189_steps4",
-        "dataset": "random",
-        "task": "i2v",
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "num-input-images": 1,
-        "seed": 42,
-        "random-request-config": [
-          {
-            "width": 1280,
-            "height": 720,
-            "prompt": "The scene comes to life with smooth, natural motion.",
-            "num_inference_steps": 4,
-            "num_frames": 189,
-            "fps": 24,
-            "weight": 1
-          }
-        ],
-        "extra-body": {
+        "dataset_name": "random-mm",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "random_range_ratio": 0.0,
+        "percentile_metrics": "e2el",
+        "random_mm_base_items_per_request": 1,
+        "random_mm_num_mm_items_range_ratio": 0,
+        "random_mm_limit_mm_per_prompt": {
+          "image": 1
+        },
+        "random_mm_bucket_config": {
+          "(720, 1280, 1)": 1.0
+        },
+        "extra_body": {
+          "width": 1280,
+          "height": 720,
+          "num_inference_steps": 4,
+          "num_frames": 189,
+          "fps": 24,
+          "seed": 42,
           "guidance_scale": 6.0,
           "flow_shift": 10.0,
           "max_sequence_length": 4096,
@@ -186,9 +197,9 @@
         },
         "baseline": {
           "H100": {
-            "throughput_qps": 0.0305,
-            "latency_mean": 33.1009,
-            "peak_memory_mb_mean": 29441.1429
+            "request_throughput": 0.0305,
+            "mean_e2el_ms": 33100.9,
+            "mean_peak_memory_mb": 29441.1429
           }
         }
       }
@@ -200,7 +211,10 @@
       {
         "hardware_marks": {
           "res": {
-            "cuda": ["H100", "B200"]
+            "cuda": [
+              "H100",
+              "B200"
+            ]
           },
           "num_cards": 2
         }
@@ -209,8 +223,6 @@
       "diffusion"
     ],
     "description": "Cosmos3-Nano video-to-video official demo workload on 2 GPUs",
-    "server_type": "vllm-omni",
-    "benchmark_endpoint": "/v1/videos",
     "server_params": {
       "model": "nvidia/Cosmos3-Nano",
       "serve_args": {
@@ -229,23 +241,30 @@
     "benchmark_params": [
       {
         "name": "1280x720_frames189_steps4",
-        "dataset": "random",
-        "task": "v2v",
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "seed": 42,
-        "random-request-config": [
-          {
-            "width": 1280,
-            "height": 720,
-            "prompt": "Continue the same scene with smooth natural motion.",
-            "num_inference_steps": 4,
-            "num_frames": 189,
-            "fps": 24,
-            "weight": 1
-          }
-        ],
-        "extra-body": {
+        "dataset_name": "random-mm",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "random_range_ratio": 0.0,
+        "percentile_metrics": "e2el",
+        "random_mm_base_items_per_request": 1,
+        "random_mm_num_mm_items_range_ratio": 0,
+        "random_mm_limit_mm_per_prompt": {
+          "video": 1
+        },
+        "random_mm_bucket_config": {
+          "(720, 1280, 189)": 1.0
+        },
+        "extra_body": {
+          "width": 1280,
+          "height": 720,
+          "num_inference_steps": 4,
+          "num_frames": 189,
+          "fps": 24,
+          "seed": 42,
           "guidance_scale": 6.0,
           "flow_shift": 10.0,
           "max_sequence_length": 4096,
@@ -258,9 +277,9 @@
         },
         "baseline": {
           "H100": {
-            "throughput_qps": 0.0299,
-            "latency_mean": 33.925,
-            "peak_memory_mb_mean": 29443.1429
+            "request_throughput": 0.0299,
+            "mean_e2el_ms": 33925.0,
+            "mean_peak_memory_mb": 29443.1429
           }
         }
       }
diff --git a/tests/dfx/perf/tests/test_hunyuanvideo15_i2v_vllm_omni.json b/tests/dfx/perf/tests/test_hunyuanvideo15_i2v_vllm_omni.json
index 25d91c44fda..f14e77bcb92 100644
--- a/tests/dfx/perf/tests/test_hunyuanvideo15_i2v_vllm_omni.json
+++ b/tests/dfx/perf/tests/test_hunyuanvideo15_i2v_vllm_omni.json
@@ -5,7 +5,10 @@
       {
         "hardware_marks": {
           "res": {
-            "cuda": ["H100", "B200"]
+            "cuda": [
+              "H100",
+              "B200"
+            ]
           },
           "num_cards": 1
         }
@@ -14,7 +17,6 @@
       "diffusion"
     ],
     "description": "Single-device baseline",
-    "server_type": "vllm-omni",
     "server_params": {
       "model": "hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_i2v",
       "serve_args": {
@@ -25,26 +27,33 @@
     "benchmark_params": [
       {
         "name": "832x480_frames33_steps4",
-        "dataset": "random",
-        "task": "i2v",
-        "num-prompts": 10,
-        "max-concurrency": 1,
-        "num-input-images": 1,
-        "seed": 42,
-        "enable-negative-prompt": true,
-        "random-request-config": [
-          {
-            "width": 832,
-            "height": 480,
-            "num_inference_steps": 4,
-            "num_frames": 33,
-            "fps": 24,
-            "weight": 1
-          }
-        ],
-        "extra-body": {
+        "dataset_name": "random-mm",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 10,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "random_range_ratio": 0.0,
+        "percentile_metrics": "e2el",
+        "random_mm_base_items_per_request": 1,
+        "random_mm_num_mm_items_range_ratio": 0,
+        "random_mm_limit_mm_per_prompt": {
+          "image": 1
+        },
+        "random_mm_bucket_config": {
+          "(480, 832, 1)": 1.0
+        },
+        "extra_body": {
+          "width": 832,
+          "height": 480,
+          "num_inference_steps": 4,
+          "num_frames": 33,
+          "fps": 24,
+          "seed": 42,
           "guidance_scale": 6.0,
-          "flow_shift": 5.0
+          "flow_shift": 5.0,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
         }
       }
     ]
@@ -55,7 +64,10 @@
       {
         "hardware_marks": {
           "res": {
-            "cuda": ["H100", "B200"]
+            "cuda": [
+              "H100",
+              "B200"
+            ]
           },
           "num_cards": 2
         }
@@ -64,7 +76,6 @@
       "diffusion"
     ],
     "description": "CacheDiT + TP=2 + VAE patch parallel=2 + VAE tiling",
-    "server_type": "vllm-omni",
     "server_params": {
       "model": "hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_i2v",
       "serve_args": {
@@ -79,26 +90,33 @@
     "benchmark_params": [
       {
         "name": "832x480_frames33_steps4",
-        "dataset": "random",
-        "task": "i2v",
-        "num-prompts": 10,
-        "max-concurrency": 1,
-        "num-input-images": 1,
-        "seed": 42,
-        "enable-negative-prompt": true,
-        "random-request-config": [
-          {
-            "width": 832,
-            "height": 480,
-            "num_inference_steps": 4,
-            "num_frames": 33,
-            "fps": 24,
-            "weight": 1
-          }
-        ],
-        "extra-body": {
+        "dataset_name": "random-mm",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 10,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "random_range_ratio": 0.0,
+        "percentile_metrics": "e2el",
+        "random_mm_base_items_per_request": 1,
+        "random_mm_num_mm_items_range_ratio": 0,
+        "random_mm_limit_mm_per_prompt": {
+          "image": 1
+        },
+        "random_mm_bucket_config": {
+          "(480, 832, 1)": 1.0
+        },
+        "extra_body": {
+          "width": 832,
+          "height": 480,
+          "num_inference_steps": 4,
+          "num_frames": 33,
+          "fps": 24,
+          "seed": 42,
           "guidance_scale": 6.0,
-          "flow_shift": 5.0
+          "flow_shift": 5.0,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
         }
       }
     ]
diff --git a/tests/dfx/perf/tests/test_hunyuanvideo15_t2v_vllm_omni.json b/tests/dfx/perf/tests/test_hunyuanvideo15_t2v_vllm_omni.json
index 6a3433dbe2c..3a2f66d82c9 100644
--- a/tests/dfx/perf/tests/test_hunyuanvideo15_t2v_vllm_omni.json
+++ b/tests/dfx/perf/tests/test_hunyuanvideo15_t2v_vllm_omni.json
@@ -5,7 +5,10 @@
       {
         "hardware_marks": {
           "res": {
-            "cuda": ["H100", "B200"]
+            "cuda": [
+              "H100",
+              "B200"
+            ]
           },
           "num_cards": 1
         }
@@ -14,7 +17,6 @@
       "diffusion"
     ],
     "description": "Single-device baseline",
-    "server_type": "vllm-omni",
     "server_params": {
       "model": "hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_t2v",
       "serve_args": {
@@ -25,25 +27,24 @@
     "benchmark_params": [
       {
         "name": "832x480_frames33_steps4",
-        "dataset": "random",
-        "task": "t2v",
-        "num-prompts": 10,
-        "max-concurrency": 1,
-        "seed": 42,
-        "enable-negative-prompt": true,
-        "random-request-config": [
-          {
-            "width": 832,
-            "height": 480,
-            "num_inference_steps": 4,
-            "num_frames": 33,
-            "fps": 24,
-            "weight": 1
-          }
-        ],
-        "extra-body": {
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 10,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 832,
+          "height": 480,
+          "num_inference_steps": 4,
+          "num_frames": 33,
+          "fps": 24,
+          "seed": 42,
           "guidance_scale": 6.0,
-          "flow_shift": 5.0
+          "flow_shift": 5.0,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
         }
       }
     ]
@@ -64,7 +65,6 @@
       "diffusion"
     ],
     "description": "CacheDiT + TP=2 + VAE patch parallel=2 + VAE tiling",
-    "server_type": "vllm-omni",
     "server_params": {
       "model": "hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_t2v",
       "serve_args": {
@@ -79,25 +79,24 @@
     "benchmark_params": [
       {
         "name": "832x480_frames33_steps4",
-        "dataset": "random",
-        "task": "t2v",
-        "num-prompts": 10,
-        "max-concurrency": 1,
-        "seed": 42,
-        "enable-negative-prompt": true,
-        "random-request-config": [
-          {
-            "width": 832,
-            "height": 480,
-            "num_inference_steps": 4,
-            "num_frames": 33,
-            "fps": 24,
-            "weight": 1
-          }
-        ],
-        "extra-body": {
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 10,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 832,
+          "height": 480,
+          "num_inference_steps": 4,
+          "num_frames": 33,
+          "fps": 24,
+          "seed": 42,
           "guidance_scale": 6.0,
-          "flow_shift": 5.0
+          "flow_shift": 5.0,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
         }
       }
     ]
diff --git a/tests/dfx/perf/tests/test_lingbot_video_vllm_omni.json b/tests/dfx/perf/tests/test_lingbot_video_vllm_omni.json
index 07403e78f9a..9f85a14a2d5 100644
--- a/tests/dfx/perf/tests/test_lingbot_video_vllm_omni.json
+++ b/tests/dfx/perf/tests/test_lingbot_video_vllm_omni.json
@@ -1,33 +1,50 @@
 [
-    {
-        "test_name": "test_lingbot_video_single_device",
-        "description": "Single-device LingBot dense T2V baseline at 320x192, 9 frames, 2 steps.",
-        "server_type": "vllm-omni",
-        "server_params": {
-            "model": "robbyant/lingbot-video-dense-1.3b",
-            "serve_args": {
-                "model-class-name": "LingBotVideoPipeline",
-                "enable-diffusion-pipeline-profiler": true
-            }
-        },
-        "benchmark_params": [
-            {
-                "name": "320x192_frames9_steps2",
-                "dataset": "random",
-                "task": "t2v",
-                "backend": "v1/videos",
-                "width": 320,
-                "height": 192,
-                "num-frames": 9,
-                "fps": 24,
-                "num-inference-steps": 2,
-                "num-prompts": 3,
-                "max-concurrency": 1,
-                "warmup-requests": 1,
-                "warmup-concurrency": 1,
-                "warmup-num-inference-steps": 2,
-                "enable-negative-prompt": true
-            }
-        ]
-    }
+  {
+    "test_name": "test_lingbot_video_single_device",
+    "mark": [
+      {
+        "hardware_marks": {
+          "res": {
+            "cuda": [
+              "H100",
+              "B200"
+            ]
+          },
+          "num_cards": 1
+        }
+      },
+      "full_model",
+      "diffusion"
+    ],
+    "description": "Single-device LingBot dense T2V baseline at 320x192, 9 frames, 2 steps.",
+    "server_params": {
+      "model": "robbyant/lingbot-video-dense-1.3b",
+      "serve_args": {
+        "model-class-name": "LingBotVideoPipeline",
+        "enable-diffusion-pipeline-profiler": true
+      }
+    },
+    "benchmark_params": [
+      {
+        "name": "320x192_frames9_steps2",
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "num_warmups": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 320,
+          "height": 192,
+          "num_frames": 9,
+          "fps": 24,
+          "num_inference_steps": 2,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
+        }
+      }
+    ]
+  }
 ]
diff --git a/tests/dfx/perf/tests/test_ltx2_vllm_omni.json b/tests/dfx/perf/tests/test_ltx2_vllm_omni.json
index 9b0634a202d..6b4c9eb702b 100644
--- a/tests/dfx/perf/tests/test_ltx2_vllm_omni.json
+++ b/tests/dfx/perf/tests/test_ltx2_vllm_omni.json
@@ -14,7 +14,6 @@
       "diffusion"
     ],
     "description": "Single-device baseline with enforce-eager (no torch.compile)",
-    "server_type": "vllm-omni",
     "server_params": {
       "model": "Lightricks/LTX-2",
       "serve_args": {
@@ -25,31 +24,41 @@
     "benchmark_params": [
       {
         "name": "256x256_145f_steps6",
-        "dataset": "random",
-        "task": "t2v",
-        "backend": "v1/videos",
-        "width": 256,
-        "height": 256,
-        "num-frames": 145,
-        "fps": 24,
-        "num-inference-steps": 6,
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "enable-negative-prompt": true
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 256,
+          "height": 256,
+          "num_frames": 145,
+          "fps": 24,
+          "num_inference_steps": 6,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
+        }
       },
       {
         "name": "480x768_41f_steps20",
-        "dataset": "random",
-        "task": "t2v",
-        "backend": "v1/videos",
-        "width": 768,
-        "height": 480,
-        "num-frames": 41,
-        "fps": 24,
-        "num-inference-steps": 20,
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "enable-negative-prompt": true
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 768,
+          "height": 480,
+          "num_frames": 41,
+          "fps": 24,
+          "num_inference_steps": 20,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
+        }
       }
     ]
   },
@@ -68,7 +77,6 @@
       "diffusion"
     ],
     "description": "Single-device with torch.compile (default, no enforce-eager)",
-    "server_type": "vllm-omni",
     "server_params": {
       "model": "Lightricks/LTX-2",
       "serve_args": {
@@ -78,31 +86,41 @@
     "benchmark_params": [
       {
         "name": "256x256_145f_steps6",
-        "dataset": "random",
-        "task": "t2v",
-        "backend": "v1/videos",
-        "width": 256,
-        "height": 256,
-        "num-frames": 145,
-        "fps": 24,
-        "num-inference-steps": 6,
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "enable-negative-prompt": true
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 256,
+          "height": 256,
+          "num_frames": 145,
+          "fps": 24,
+          "num_inference_steps": 6,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
+        }
       },
       {
         "name": "480x768_41f_steps20",
-        "dataset": "random",
-        "task": "t2v",
-        "backend": "v1/videos",
-        "width": 768,
-        "height": 480,
-        "num-frames": 41,
-        "fps": 24,
-        "num-inference-steps": 20,
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "enable-negative-prompt": true
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 768,
+          "height": 480,
+          "num_frames": 41,
+          "fps": 24,
+          "num_inference_steps": 20,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
+        }
       }
     ]
   },
@@ -121,7 +139,6 @@
       "diffusion"
     ],
     "description": "CFG-parallel=2 with enforce-eager",
-    "server_type": "vllm-omni",
     "server_params": {
       "model": "Lightricks/LTX-2",
       "serve_args": {
@@ -133,31 +150,41 @@
     "benchmark_params": [
       {
         "name": "256x256_145f_steps6",
-        "dataset": "random",
-        "task": "t2v",
-        "backend": "v1/videos",
-        "width": 256,
-        "height": 256,
-        "num-frames": 145,
-        "fps": 24,
-        "num-inference-steps": 6,
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "enable-negative-prompt": true
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 256,
+          "height": 256,
+          "num_frames": 145,
+          "fps": 24,
+          "num_inference_steps": 6,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
+        }
       },
       {
         "name": "480x768_41f_steps20",
-        "dataset": "random",
-        "task": "t2v",
-        "backend": "v1/videos",
-        "width": 768,
-        "height": 480,
-        "num-frames": 41,
-        "fps": 24,
-        "num-inference-steps": 20,
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "enable-negative-prompt": true
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 768,
+          "height": 480,
+          "num_frames": 41,
+          "fps": 24,
+          "num_inference_steps": 20,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
+        }
       }
     ]
   },
@@ -176,7 +203,6 @@
       "diffusion"
     ],
     "description": "CFG-parallel=2 with torch.compile",
-    "server_type": "vllm-omni",
     "server_params": {
       "model": "Lightricks/LTX-2",
       "serve_args": {
@@ -187,31 +213,41 @@
     "benchmark_params": [
       {
         "name": "256x256_145f_steps6",
-        "dataset": "random",
-        "task": "t2v",
-        "backend": "v1/videos",
-        "width": 256,
-        "height": 256,
-        "num-frames": 145,
-        "fps": 24,
-        "num-inference-steps": 6,
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "enable-negative-prompt": true
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 256,
+          "height": 256,
+          "num_frames": 145,
+          "fps": 24,
+          "num_inference_steps": 6,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
+        }
       },
       {
         "name": "480x768_41f_steps20",
-        "dataset": "random",
-        "task": "t2v",
-        "backend": "v1/videos",
-        "width": 768,
-        "height": 480,
-        "num-frames": 41,
-        "fps": 24,
-        "num-inference-steps": 20,
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "enable-negative-prompt": true
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 768,
+          "height": 480,
+          "num_frames": 41,
+          "fps": 24,
+          "num_inference_steps": 20,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
+        }
       }
     ]
   },
@@ -230,7 +266,6 @@
       "diffusion"
     ],
     "description": "CacheDiT with enforce-eager",
-    "server_type": "vllm-omni",
     "server_params": {
       "model": "Lightricks/LTX-2",
       "serve_args": {
@@ -242,31 +277,41 @@
     "benchmark_params": [
       {
         "name": "256x256_145f_steps6",
-        "dataset": "random",
-        "task": "t2v",
-        "backend": "v1/videos",
-        "width": 256,
-        "height": 256,
-        "num-frames": 145,
-        "fps": 24,
-        "num-inference-steps": 6,
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "enable-negative-prompt": true
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 256,
+          "height": 256,
+          "num_frames": 145,
+          "fps": 24,
+          "num_inference_steps": 6,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
+        }
       },
       {
         "name": "480x768_41f_steps20",
-        "dataset": "random",
-        "task": "t2v",
-        "backend": "v1/videos",
-        "width": 768,
-        "height": 480,
-        "num-frames": 41,
-        "fps": 24,
-        "num-inference-steps": 20,
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "enable-negative-prompt": true
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 768,
+          "height": 480,
+          "num_frames": 41,
+          "fps": 24,
+          "num_inference_steps": 20,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
+        }
       }
     ]
   }
diff --git a/tests/dfx/perf/tests/test_minimax_h3_vllm_omni.json b/tests/dfx/perf/tests/test_minimax_h3_vllm_omni.json
index 7dfbe6f3de9..3bf68064eee 100644
--- a/tests/dfx/perf/tests/test_minimax_h3_vllm_omni.json
+++ b/tests/dfx/perf/tests/test_minimax_h3_vllm_omni.json
@@ -5,7 +5,10 @@
       {
         "hardware_marks": {
           "res": {
-            "cuda": ["H100", "B200"]
+            "cuda": [
+              "H100",
+              "B200"
+            ]
           },
           "num_cards": 4
         }
@@ -14,8 +17,6 @@
       "diffusion"
     ],
     "description": "MiniMax-H3 T2VA text-to-video+audio on 4 GPUs (USP4 + HSDP4, encoder-TP4, VAE-PP4 tile)",
-    "server_type": "vllm-omni",
-    "benchmark_endpoint": "/v1/videos",
     "server_params": {
       "model": "MiniMaxAI/MiniMax-H3",
       "serve_args": {
@@ -38,33 +39,30 @@
     "benchmark_params": [
       {
         "name": "1344x768_frames209_steps8",
-        "dataset": "random",
-        "task": "t2v",
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "warmup-requests": 1,
-        "warmup-concurrency": 1,
-        "seed": 42,
-        "random-request-config": [
-          {
-            "width": 1344,
-            "height": 768,
-            "prompt": "Cinematic drone shot pushing through misty mountain ridges at sunrise, golden light catching the clouds, slow and smooth.",
-            "num_inference_steps": 8,
-            "num_frames": 209,
-            "fps": 24,
-            "weight": 1
-          }
-        ],
-        "extra-body": {
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "num_warmups": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 1344,
+          "height": 768,
+          "num_inference_steps": 8,
+          "num_frames": 209,
+          "fps": 24,
+          "seed": 42,
           "flow_shift": 12.0,
           "extra_params": "{\"task\":\"t2va\",\"duration\":8.7,\"aspect_ratio\":\"16:9\",\"audio_flow_shift\":3.0}"
         },
         "baseline": {
           "H100": {
-            "throughput_qps": 0.0211,
-            "latency_mean": 38.3252,
-            "peak_memory_mb_mean": 54219.6
+            "request_throughput": 0.0211,
+            "mean_e2el_ms": 38325.2,
+            "mean_peak_memory_mb": 54219.6
           }
         }
       }
@@ -76,7 +74,10 @@
       {
         "hardware_marks": {
           "res": {
-            "cuda": ["H100", "B200"]
+            "cuda": [
+              "H100",
+              "B200"
+            ]
           },
           "num_cards": 4
         }
@@ -85,8 +86,6 @@
       "diffusion"
     ],
     "description": "MiniMax-H3 FL2VA first-frame image-to-video+audio on 4 GPUs (USP4 + HSDP4, encoder-TP4, VAE-PP4 tile)",
-    "server_type": "vllm-omni",
-    "benchmark_endpoint": "/v1/videos",
     "server_params": {
       "model": "MiniMaxAI/MiniMax-H3",
       "serve_args": {
@@ -109,34 +108,39 @@
     "benchmark_params": [
       {
         "name": "1344x768_frames209_steps8",
-        "dataset": "random",
-        "task": "ti2v",
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "warmup-requests": 1,
-        "warmup-concurrency": 1,
-        "seed": 42,
-        "random-request-config": [
-          {
-            "width": 1344,
-            "height": 768,
-            "prompt": "Continue the scene: the camera slowly pushes in on the subject as the background city lights drift and blur, with the actor singing softly and emotionally.",
-            "num_inference_steps": 8,
-            "num_frames": 209,
-            "fps": 24,
-            "weight": 1
-          }
-        ],
-        "extra-body": {
+        "dataset_name": "random-mm",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "num_warmups": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "random_range_ratio": 0.0,
+        "percentile_metrics": "e2el",
+        "random_mm_base_items_per_request": 1,
+        "random_mm_num_mm_items_range_ratio": 0,
+        "random_mm_limit_mm_per_prompt": {
+          "image": 1
+        },
+        "random_mm_bucket_config": {
+          "(768, 1344, 1)": 1.0
+        },
+        "extra_body": {
+          "width": 1344,
+          "height": 768,
+          "num_inference_steps": 8,
+          "num_frames": 209,
+          "fps": 24,
+          "seed": 42,
           "flow_shift": 12.0,
           "extra_params": "{\"task\":\"fl2va\",\"duration\":8.7,\"audio_flow_shift\":3.0}"
         },
-        "num-input-images": 1,
         "baseline": {
           "H100": {
-            "throughput_qps": 0.0212,
-            "latency_mean": 38.1385,
-            "peak_memory_mb_mean": 54622.8
+            "request_throughput": 0.0212,
+            "mean_e2el_ms": 38138.5,
+            "mean_peak_memory_mb": 54622.8
           }
         }
       }
@@ -148,7 +152,10 @@
       {
         "hardware_marks": {
           "res": {
-            "cuda": ["H100", "B200"]
+            "cuda": [
+              "H100",
+              "B200"
+            ]
           },
           "num_cards": 4
         }
@@ -157,8 +164,6 @@
       "diffusion"
     ],
     "description": "MiniMax-H3 Ref2VA video-reference-to-video+audio on 4 GPUs (USP4 + HSDP4, encoder-TP4, VAE-PP4 tile)",
-    "server_type": "vllm-omni",
-    "benchmark_endpoint": "/v1/videos",
     "server_params": {
       "model": "MiniMaxAI/MiniMax-H3",
       "serve_args": {
@@ -181,33 +186,39 @@
     "benchmark_params": [
       {
         "name": "1344x768_frames209_steps8",
-        "dataset": "random",
-        "task": "v2v",
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "warmup-requests": 1,
-        "warmup-concurrency": 1,
-        "seed": 42,
-        "random-request-config": [
-          {
-            "width": 1344,
-            "height": 768,
-            "prompt": "Restyle the reference footage into a cinematic night scene with neon reflections, keeping the same camera motion and subject.",
-            "num_inference_steps": 8,
-            "num_frames": 209,
-            "fps": 24,
-            "weight": 1
-          }
-        ],
-        "extra-body": {
+        "dataset_name": "random-mm",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "num_warmups": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "random_range_ratio": 0.0,
+        "percentile_metrics": "e2el",
+        "random_mm_base_items_per_request": 1,
+        "random_mm_num_mm_items_range_ratio": 0,
+        "random_mm_limit_mm_per_prompt": {
+          "video": 1
+        },
+        "random_mm_bucket_config": {
+          "(480, 832, 209)": 1.0
+        },
+        "extra_body": {
+          "width": 1344,
+          "height": 768,
+          "num_inference_steps": 8,
+          "num_frames": 209,
+          "fps": 24,
+          "seed": 42,
           "flow_shift": 12.0,
           "extra_params": "{\"task\":\"ref2va\",\"duration\":8.7,\"aspect_ratio\":\"16:9\",\"audio_flow_shift\":3.0}"
         },
         "baseline": {
           "H100": {
-            "throughput_qps": 0.0077,
-            "latency_mean": 105.7508,
-            "peak_memory_mb_mean": 55180.8
+            "request_throughput": 0.0077,
+            "mean_e2el_ms": 105750.8,
+            "mean_peak_memory_mb": 55180.8
           }
         }
       }
@@ -219,7 +230,10 @@
       {
         "hardware_marks": {
           "res": {
-            "cuda": ["H100", "B200"]
+            "cuda": [
+              "H100",
+              "B200"
+            ]
           },
           "num_cards": 4
         }
@@ -228,8 +242,6 @@
       "diffusion"
     ],
     "description": "MiniMax-H3 T2VA with 4-card distributed layerwise offload (DLO), 8 denoise steps (USP4 + HSDP4, encoder-TP4, VAE-PP4 tile)",
-    "server_type": "vllm-omni",
-    "benchmark_endpoint": "/v1/videos",
     "server_params": {
       "model": "MiniMaxAI/MiniMax-H3",
       "serve_args": {
@@ -256,33 +268,30 @@
     "benchmark_params": [
       {
         "name": "1344x768_frames209_steps8",
-        "dataset": "random",
-        "task": "t2v",
-        "num-prompts": 3,
-        "max-concurrency": 1,
-        "warmup-requests": 1,
-        "warmup-concurrency": 1,
-        "seed": 42,
-        "random-request-config": [
-          {
-            "width": 1344,
-            "height": 768,
-            "prompt": "Cinematic drone shot pushing through misty mountain ridges at sunrise, golden light catching the clouds, slow and smooth.",
-            "num_inference_steps": 8,
-            "num_frames": 209,
-            "fps": 24,
-            "weight": 1
-          }
-        ],
-        "extra-body": {
+        "dataset_name": "random",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 3,
+        "max_concurrency": 1,
+        "num_warmups": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "percentile_metrics": "e2el",
+        "extra_body": {
+          "width": 1344,
+          "height": 768,
+          "num_inference_steps": 8,
+          "num_frames": 209,
+          "fps": 24,
+          "seed": 42,
           "flow_shift": 12.0,
           "extra_params": "{\"task\":\"t2va\",\"duration\":8.7,\"aspect_ratio\":\"16:9\",\"audio_flow_shift\":3.0}"
         },
         "baseline": {
           "H100": {
-            "throughput_qps": 0.0212,
-            "latency_mean": 38.2089,
-            "peak_memory_mb_mean": 36218.4
+            "request_throughput": 0.0212,
+            "mean_e2el_ms": 38208.9,
+            "mean_peak_memory_mb": 36218.4
           }
         }
       }
diff --git a/tests/dfx/perf/tests/test_runner_metadata.py b/tests/dfx/perf/tests/test_runner_metadata.py
index a0f966121a1..bdbd83df0a3 100644
--- a/tests/dfx/perf/tests/test_runner_metadata.py
+++ b/tests/dfx/perf/tests/test_runner_metadata.py
@@ -205,15 +205,53 @@ def test_is_diffusion_perf_config():
     from tests.dfx.conftest import is_diffusion_perf_config
 
     assert not is_diffusion_perf_config(
-        {"test_name": "omni_a", "mark": [{"hardware_marks": {"res": {"cuda": "H100"}}}, "omni"]}
+        {
+            "test_name": "omni_a",
+            "mark": [{"hardware_marks": {"res": {"cuda": "H100"}}}, "omni"],
+            "benchmark_params": [{"dataset_name": "random", "endpoint": "/v1/chat/completions"}],
+        }
     )
     assert is_diffusion_perf_config(
         {
             "test_name": "diff_a",
             "server_type": "vllm-omni",
             "mark": [{"hardware_marks": {"res": {"cuda": "H100"}}}, "diffusion"],
+            "benchmark_params": [{"task": "t2i", "dataset": "random"}],
         }
     )
+    videos_cfg = {
+        "test_name": "diff_videos",
+        "server_type": "vllm-omni",
+        "mark": [{"hardware_marks": {"res": {"cuda": "H100"}}}, "diffusion"],
+        "benchmark_params": [{"task": "t2v", "dataset_name": "random", "endpoint": "/v1/videos"}],
+    }
+    assert not is_diffusion_perf_config(videos_cfg)
+    custom_edits_cfg = {
+        "test_name": "diff_custom_edits",
+        "server_type": "vllm-omni",
+        "benchmark_endpoint": "/v1/images/edits",
+        "benchmark_params": [{"dataset": "custom", "task": "ti2i"}],
+    }
+    assert is_diffusion_perf_config(custom_edits_cfg)
+
+
+def test_merge_omni_default_server_args_respects_json():
+    from tests.dfx.perf.scripts.run_benchmark import _merge_omni_default_server_args
+
+    extra = ("--stage-init-timeout", "1800", "--init-timeout", "1800", "--usp", "4")
+    assert _merge_omni_default_server_args(extra, use_omni=True) == []
+    assert _merge_omni_default_server_args((), use_omni=True) == [
+        "--stage-init-timeout",
+        "600",
+        "--init-timeout",
+        "900",
+    ]
+    assert _merge_omni_default_server_args((), use_omni=False) == []
+    # Only fill the missing default; keep JSON's other timeouts.
+    assert _merge_omni_default_server_args(("--stage-init-timeout=1800",), use_omni=True) == [
+        "--init-timeout",
+        "900",
+    ]
 
 
 def test_benchmark_param_id_suffix_from_task_eval_phase():
@@ -303,12 +341,84 @@ def fake_start(server_param):
     finally:
         active_context.close()
 
-    assert events == [
-        ("start", omni_p0_server),
-        ("stop", omni_p0_server),
-        ("start", tts_server),
-        ("stop", tts_server),
-    ]
+
+def test_run_benchmark_persists_distinct_benchmark_params_name(tmp_path, monkeypatch):
+    """Two benchmark_params under one test_name must keep distinct saved identity."""
+    import io
+    from pathlib import Path
+
+    from tests.dfx import conftest as dfx_conftest
+
+    result_dir = tmp_path / "results"
+    result_dir.mkdir()
+    monkeypatch.setenv("BENCHMARK_DIR", str(result_dir))
+
+    class _FakePopen:
+        def __init__(self, command, **kwargs):
+            self.stdout = io.StringIO("")
+            self.stderr = io.StringIO("")
+            result_dir_idx = command.index("--result-dir")
+            filename_idx = command.index("--result-filename")
+            out = Path(command[result_dir_idx + 1]) / command[filename_idx + 1]
+            out.parent.mkdir(parents=True, exist_ok=True)
+            out.write_text(json.dumps({"completed": 3, "request_throughput": 0.01}), encoding="utf-8")
+
+        def wait(self):
+            return 0
+
+    monkeypatch.setattr(dfx_conftest.subprocess, "Popen", _FakePopen)
+
+    shared_test = "test_wan22_i2v_usp2"
+    name_a = "832x480_frames81_steps4"
+    name_b = "1280x720_frames121_steps4"
+
+    result_a = dfx_conftest.run_benchmark(
+        args=["--host", "127.0.0.1", "--port", "8000"],
+        test_name=shared_test,
+        flow=1,
+        dataset_name="random-mm",
+        num_prompt=10,
+        random_input_len=8,
+        random_output_len=1,
+        resource_label="H800",
+        benchmark_params_name=name_a,
+    )
+    result_b = dfx_conftest.run_benchmark(
+        args=["--host", "127.0.0.1", "--port", "8000"],
+        test_name=shared_test,
+        flow=1,
+        dataset_name="random-mm",
+        num_prompt=10,
+        random_input_len=8,
+        random_output_len=1,
+        resource_label="H800",
+        benchmark_params_name=name_b,
+    )
+
+    assert result_a["test_name"] == shared_test
+    assert result_b["test_name"] == shared_test
+    assert result_a["name"] == name_a
+    assert result_b["name"] == name_b
+    assert "benchmark_params" not in result_a
+    assert "benchmark_params" not in result_b
+
+    files = sorted(p.name for p in result_dir.glob("result_*.json"))
+    assert len(files) == 2
+    assert any(name_a in name for name in files)
+    assert any(name_b in name for name in files)
+    assert files[0] != files[1]
+
+    def omni_group_key(record: dict) -> tuple:
+        return (
+            record.get("model_id") or "",
+            record.get("test_name") or "",
+            record.get("name") or "",
+            record.get("dataset_name") or "",
+            record.get("max_concurrency") if record.get("max_concurrency") is not None else 0,
+            record.get("num_prompts") if record.get("num_prompts") is not None else 0,
+        )
+
+    assert omni_group_key(result_a) != omni_group_key(result_b)
 
 
 def test_is_hardware_nested_baseline():
diff --git a/tests/dfx/perf/tests/test_wan22_i2v_vllm_omni.json b/tests/dfx/perf/tests/test_wan22_i2v_vllm_omni.json
index 3244d1c4155..0aa060b4cbc 100644
--- a/tests/dfx/perf/tests/test_wan22_i2v_vllm_omni.json
+++ b/tests/dfx/perf/tests/test_wan22_i2v_vllm_omni.json
@@ -5,7 +5,10 @@
       {
         "hardware_marks": {
           "res": {
-            "cuda": ["H100", "B200"]
+            "cuda": [
+              "H100",
+              "B200"
+            ]
           },
           "num_cards": 1
         }
@@ -14,7 +17,6 @@
       "diffusion"
     ],
     "description": "Single-device baseline",
-    "server_type": "vllm-omni",
     "server_params": {
       "model": "Wan-AI/Wan2.2-I2V-A14B-Diffusers",
       "serve_args": {
@@ -24,28 +26,37 @@
     "benchmark_params": [
       {
         "name": "832x480_frames81_steps4",
-        "dataset": "random",
-        "task": "i2v",
-        "num-prompts": 10,
-        "max-concurrency": 1,
-        "num-input-images": 1,
-        "seed": 42,
-        "enable-negative-prompt": true,
-        "random-request-config": [
-          {
-            "width": 832,
-            "height": 480,
-            "num_inference_steps": 4,
-            "num_frames": 81,
-            "fps": 16,
-            "weight": 1
-          }
-        ],
+        "dataset_name": "random-mm",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 10,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "random_range_ratio": 0.0,
+        "percentile_metrics": "e2el",
+        "random_mm_base_items_per_request": 1,
+        "random_mm_num_mm_items_range_ratio": 0,
+        "random_mm_limit_mm_per_prompt": {
+          "image": 1
+        },
+        "random_mm_bucket_config": {
+          "(480, 832, 1)": 1.0
+        },
+        "extra_body": {
+          "width": 832,
+          "height": 480,
+          "num_inference_steps": 4,
+          "num_frames": 81,
+          "fps": 16,
+          "seed": 42,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
+        },
         "baseline": {
           "H100": {
-            "throughput_qps": 0.0361,
-            "latency_mean": 27.6206,
-            "peak_memory_mb_mean": 80548.0
+            "request_throughput": 0.0361,
+            "mean_e2el_ms": 27620.6,
+            "mean_peak_memory_mb": 80548.0
           }
         }
       }
@@ -57,7 +68,10 @@
       {
         "hardware_marks": {
           "res": {
-            "cuda": ["H100", "B200"],
+            "cuda": [
+              "H100",
+              "B200"
+            ],
             "npu": "A3"
           },
           "num_cards": 2
@@ -67,7 +81,6 @@
       "diffusion"
     ],
     "description": "USP=2 + VAE patch parallel=2 + HSDP + VAE slicing",
-    "server_type": "vllm-omni",
     "server_params": {
       "model": "Wan-AI/Wan2.2-I2V-A14B-Diffusers",
       "serve_args": {
@@ -81,55 +94,73 @@
     "benchmark_params": [
       {
         "name": "832x480_frames81_steps4",
-        "dataset": "random",
-        "task": "i2v",
-        "num-prompts": 10,
-        "max-concurrency": 1,
-        "num-input-images": 1,
-        "seed": 42,
-        "enable-negative-prompt": true,
-        "random-request-config": [
-          {
-            "width": 832,
-            "height": 480,
-            "num_inference_steps": 4,
-            "num_frames": 81,
-            "fps": 16,
-            "weight": 1
-          }
-        ],
+        "dataset_name": "random-mm",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 10,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "random_range_ratio": 0.0,
+        "percentile_metrics": "e2el",
+        "random_mm_base_items_per_request": 1,
+        "random_mm_num_mm_items_range_ratio": 0,
+        "random_mm_limit_mm_per_prompt": {
+          "image": 1
+        },
+        "random_mm_bucket_config": {
+          "(480, 832, 1)": 1.0
+        },
+        "extra_body": {
+          "width": 832,
+          "height": 480,
+          "num_inference_steps": 4,
+          "num_frames": 81,
+          "fps": 16,
+          "seed": 42,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
+        },
         "baseline": {
           "H100": {
-            "throughput_qps": 0.05263,
-            "latency_mean": 18.9415,
-            "peak_memory_mb_mean": 50053.7
+            "request_throughput": 0.05263,
+            "mean_e2el_ms": 18941.5,
+            "mean_peak_memory_mb": 50053.7
           }
         }
       },
       {
         "name": "1280x720_frames121_steps4",
-        "dataset": "random",
-        "task": "i2v",
-        "num-prompts": 10,
-        "max-concurrency": 1,
-        "num-input-images": 1,
-        "seed": 42,
-        "enable-negative-prompt": true,
-        "random-request-config": [
-          {
-            "width": 1280,
-            "height": 720,
-            "num_inference_steps": 4,
-            "num_frames": 121,
-            "fps": 16,
-            "weight": 1
-          }
-        ],
+        "dataset_name": "random-mm",
+        "tokenizer": "gpt2",
+        "endpoint": "/v1/videos",
+        "num_prompts": 10,
+        "max_concurrency": 1,
+        "random_input_len": 8,
+        "random_output_len": 1,
+        "random_range_ratio": 0.0,
+        "percentile_metrics": "e2el",
+        "random_mm_base_items_per_request": 1,
+        "random_mm_num_mm_items_range_ratio": 0,
+        "random_mm_limit_mm_per_prompt": {
+          "image": 1
+        },
+        "random_mm_bucket_config": {
+          "(720, 1280, 1)": 1.0
+        },
+        "extra_body": {
+          "width": 1280,
+          "height": 720,
+          "num_inference_steps": 4,
+          "num_frames": 121,
+          "fps": 16,
+          "seed": 42,
+          "negative_prompt": "Negative prompt for benchmarking diffusion models"
+        },
         "baseline": {
           "H100": {
-            "throughput_qps": 0.009595,
-            "latency_mean": 104.1058,
-            "peak_memory_mb_mean": 59718.0
+            "request_throughput": 0.009595,
+            "mean_e2el_ms": 104105.8,
+            "mean_peak_memory_mb": 59718.0
           }
         }
       }
diff --git a/tools/nightly/generate_nightly_perf_excel.py b/tools/nightly/generate_nightly_perf_excel.py
index 54c4b3f3406..eae463173a0 100644
--- a/tools/nightly/generate_nightly_perf_excel.py
+++ b/tools/nightly/generate_nightly_perf_excel.py
@@ -1,4 +1,7 @@
 #!/usr/bin/env python3
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM-Omni project
+
 """
 Generate a nightly Excel performance report from JSON results.
 
@@ -137,6 +140,7 @@ def _omni_group_key(record: dict[str, Any]) -> tuple[Any, ...]:
     return (
         record.get("model_id") or "",
         record.get("test_name") or "",
+        record.get("name") or "",
         record.get("dataset_name") or "",
         record.get("max_concurrency") if record.get("max_concurrency") is not None else 0,
         record.get("num_prompts") if record.get("num_prompts") is not None else 0,
@@ -171,6 +175,7 @@ def _load_summary_columns(script_dir: str) -> list[str]:
         "model_id",
         "tokenizer_id",
         "test_name",
+        "name",
         "dataset_name",
         "num_prompts",
         "request_rate",
@@ -218,7 +223,7 @@ def _load_summary_columns(script_dir: str) -> list[str]:
 
 def _ensure_omni_summary_columns(summary_columns: list[str]) -> list[str]:
     """Ensure omni summary contains required columns, even when a custom columns file exists."""
-    required = ("test_name", "dataset_name", "source_file")
+    required = ("test_name", "name", "dataset_name", "source_file")
     existing = set(summary_columns)
     if all(c in existing for c in required):
         return summary_columns
@@ -232,8 +237,18 @@ def _ensure_omni_summary_columns(summary_columns: list[str]) -> list[str]:
         else:
             out.append("test_name")
     existing = set(out)
-    if "dataset_name" not in existing:
+    if "name" not in existing:
         if "test_name" in out:
+            idx = out.index("test_name") + 1
+            out.insert(idx, "name")
+        else:
+            out.append("name")
+    existing = set(out)
+    if "dataset_name" not in existing:
+        if "name" in out:
+            idx = out.index("name") + 1
+            out.insert(idx, "dataset_name")
+        elif "test_name" in out:
             idx = out.index("test_name") + 1
             out.insert(idx, "dataset_name")
         else:
diff --git a/tools/nightly/generate_nightly_perf_html.py b/tools/nightly/generate_nightly_perf_html.py
index 11e9b0961ee..cb29f627c57 100644
--- a/tools/nightly/generate_nightly_perf_html.py
+++ b/tools/nightly/generate_nightly_perf_html.py
@@ -1,4 +1,7 @@
 #!/usr/bin/env python3
+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM-Omni project
+
 """
 Generate an improved nightly HTML performance dashboard without modifying the
 existing generator.
@@ -256,6 +259,7 @@ def _sort_omni_records(records: list[dict[str, Any]]) -> list[dict[str, Any]]:
         key=lambda r: (
             r.get("model_id") or "",
             r.get("test_name") or "",
+            r.get("name") or "",
             r.get("dataset_name") or "",
             r.get("max_concurrency") or 0,
             r.get("num_prompts") or 0,
@@ -1236,7 +1240,7 @@ def _build_html_document(
   const extraKey = prefix === "diff" ? "" : "tokenizer_id";
   const metaKeys = prefix === "diff"
     ? ["test_name"]
-    : ["test_name", "max_concurrency", "num_prompts"];
+    : ["test_name", "name", "max_concurrency", "num_prompts"];
   const configFields = prefix === "diff"
     ? ["test_name", "model", "endpoint", "dataset"]
     : [
@@ -1245,13 +1249,14 @@ def _build_html_document(
         "model_id",
         "tokenizer_id",
         "test_name",
+        "name",
         "dataset_name",
         "max_concurrency",
         "num_prompts",
       ];
   const labelFields = prefix === "diff"
     ? ["test_name", "dataset"]
-    : ["test_name", "dataset_name"];
+    : ["test_name", "name", "dataset_name"];
   const latestMetric = prefix === "diff" ? "throughput_qps" : "output_throughput";
   const secondaryMetric = prefix === "diff" ? "latency_mean" : "mean_e2el_ms";
   const modelList = document.getElementById(`${{prefix}}-model-list`);
@@ -1650,7 +1655,7 @@ def _build_html_document(
                 '<div class="section-subtitle">Chat/audio benchmark history with '
                 "trend charts grouped by throughput, latency, and audio metrics."
                 '</div></div><div class="section-pills"><div class="pill"><strong>'
-                "Grouping</strong> test_name + dataset + concurrency + prompts</div>"
+                "Grouping</strong> test_name + name + dataset + concurrency + prompts</div>"
                 '<div class="pill"><strong>Snapshots</strong> single-point series '
                 "stay readable</div></div></div>"
             ),
diff --git a/tools/nightly/run_nightly_jobs.sh b/tools/nightly/run_nightly_jobs.sh
index 8a08097260b..1b37740ce01 100644
--- a/tools/nightly/run_nightly_jobs.sh
+++ b/tools/nightly/run_nightly_jobs.sh
@@ -41,7 +41,7 @@
 #     From repo root: pytest -sv -m "<markers> and local_model" (markers from MODEL_TYPE: omni, tts,
 #     diffusion; all → "(omni or tts or diffusion) and local_model"). Not filtered by nightly YAML.
 #     LABEL_SUBSTR: if set, restrict to tests/**/test_*.py basenames and tests/dfx/perf/tests/*.json
-#                   whose filename contains the substring (benchmark runner chosen by JSON family).
+#                   whose filename contains the substring (benchmark runner via is_diffusion_perf_config).
 #
 #   stability (when included in TEST_TYPE):
 #     From repo root: pytest -s -v --run-level full_model -m "<mark>" tests/dfx/stability/scripts/...
@@ -711,8 +711,19 @@ def perf_json_model_family(json_basename: str) -> str:
 
 
 def perf_json_runner(json_basename: str) -> Path:
-    if perf_json_model_family(json_basename) in ("omni", "tts"):
+    """Pick benchmark runner from JSON schema (``dataset`` vs ``dataset_name``)."""
+    json_path = REPO_ROOT / PERF_TESTS_REL / json_basename
+    try:
+        if str(REPO_ROOT) not in sys.path:
+            sys.path.insert(0, str(REPO_ROOT))
+        from tests.dfx.conftest import is_diffusion_perf_config, load_configs
+
+        configs = load_configs(str(json_path))
+        if configs and any(is_diffusion_perf_config(cfg) for cfg in configs):
+            return RUN_DIFFUSION_BENCHMARK_REL
         return RUN_BENCHMARK_REL
+    except Exception as exc:
+        print(f"# warn: could not classify {json_basename}: {exc}", file=sys.stderr)
     return RUN_DIFFUSION_BENCHMARK_REL
 
 
diff --git a/vllm_omni/benchmarks/metrics/metrics.py b/vllm_omni/benchmarks/metrics/metrics.py
index 759e606671f..6cfe473d5b1 100644
--- a/vllm_omni/benchmarks/metrics/metrics.py
+++ b/vllm_omni/benchmarks/metrics/metrics.py
@@ -5,6 +5,7 @@
 from collections import defaultdict
 from collections.abc import Sequence
 from dataclasses import field, make_dataclass
+from typing import Any
 
 import numpy as np
 from vllm.benchmarks.datasets import SampleRequest
@@ -15,6 +16,7 @@
 from vllm_omni.metrics import definitions as defs
 
 _PERCENTILE_ROWS_TYPE = list[tuple[float, float]] | None
+_STAGE_DURATION_MAP_TYPE = dict[str, float] | None
 _FLOAT_LIST_TYPE = list[float]
 _INT_LIST_TYPE = list[int]
 
@@ -67,9 +69,14 @@
     (defs.MEDIAN_PEAK_MEMORY_MB, float, field(default=0.0)),
     (defs.STD_PEAK_MEMORY_MB, float, field(default=0.0)),
     (defs.PERCENTILES_PEAK_MEMORY_MB, _PERCENTILE_ROWS_TYPE, field(default=None)),
+    (defs.STAGE_DURATIONS_MEAN, _STAGE_DURATION_MAP_TYPE, field(default=None)),
+    (defs.STAGE_DURATIONS_P50, _STAGE_DURATION_MAP_TYPE, field(default=None)),
+    (defs.STAGE_DURATIONS_P99, _STAGE_DURATION_MAP_TYPE, field(default=None)),
 ]
 
-MultiModalsBenchmarkMetrics = make_dataclass(
+# ``make_dataclass`` returns a runtime class that mypy treats as a variable, not
+# a type. Annotate as Any so call sites can use these names in annotations.
+MultiModalsBenchmarkMetrics: Any = make_dataclass(
     "MultiModalsBenchmarkMetrics",
     _MULTIMODAL_BENCHMARK_FIELDS,
     bases=(BenchmarkMetrics,),
@@ -99,7 +106,7 @@
     (defs.INTER_OUTPUT_LATENCIES_MS, _FLOAT_LIST_TYPE, field(default_factory=list)),
 ]
 
-StageBenchmarkMetrics = make_dataclass(
+StageBenchmarkMetrics: Any = make_dataclass(
     "StageBenchmarkMetrics",
     _STAGE_BENCHMARK_FIELDS,
     namespace={"__doc__": "Aggregated metrics for one pipeline stage (for printing only)."},
@@ -240,6 +247,7 @@ def print_metrics(
     if isinstance(metrics, MultiModalsBenchmarkMetrics):
         print("{:<40} {:<10.2f}".format("Peak concurrent requests:", metrics.max_concurrent_requests))
         print_peak_memory_metrics(metrics)
+        print_stage_durations_metrics(selected_percentiles or [], metrics)
     if task_type != TaskType.GENERATION or "e2el" in selected_percentile_metrics:
         process_one_metric("e2el", metrics)
     print_text_metrics(task_type, selected_percentile_metrics, metrics)
@@ -338,6 +346,92 @@ def print_peak_memory_metrics(metrics: MultiModalsBenchmarkMetrics):
         print("{:<40} {:<10.2f}".format(f"P{_p_label(p)} PEAK_MEMORY_MB (MB):", value))
 
 
+def aggregate_stage_durations(outputs: Sequence[RequestFuncOutput]) -> dict[str, dict[str, float]]:
+    """Aggregate pipeline profiler timings into mean/p50/p99 maps.
+
+    ``stage_N_gen_ms`` is omitted. That clock already belongs to
+    ``stage_gen_time`` / image / video generation, not to this profiler block.
+    The maps are stored on ``MultiModalsBenchmarkMetrics`` and printed with the
+    other run-level metrics. ``--print-stage`` does not gate them.
+    """
+    stage_duration_lists: dict[str, list[float]] = {}
+    for output in outputs:
+        if not getattr(output, "success", False):
+            continue
+        stage_durations = getattr(output, "stage_durations", None)
+        if not isinstance(stage_durations, dict):
+            continue
+        for stage, duration in stage_durations.items():
+            if _is_printed_stage_gen_key(str(stage)):
+                continue
+            if isinstance(duration, bool) or not isinstance(duration, (int, float)) or not np.isfinite(duration):
+                continue
+            stage_duration_lists.setdefault(str(stage), []).append(float(duration))
+    if not stage_duration_lists:
+        return {}
+    return {
+        defs.STAGE_DURATIONS_MEAN: {stage: float(np.mean(values)) for stage, values in stage_duration_lists.items()},
+        defs.STAGE_DURATIONS_P50: {
+            stage: float(np.percentile(values, 50)) for stage, values in stage_duration_lists.items()
+        },
+        defs.STAGE_DURATIONS_P99: {
+            stage: float(np.percentile(values, 99)) for stage, values in stage_duration_lists.items()
+        },
+    }
+
+
+def _stage_duration_display_name(stage: str) -> str:
+    """Strip pipeline-class prefix and prettify profiler keys for printing.
+
+    ``Wan22I2VPipeline.diffuse`` -> ``Diffuse``
+    ``Wan22I2VPipeline.text_encoder.forward`` -> ``Text Encoder Forward``
+    ``queue_wait_ms`` -> ``Queue Wait``
+    """
+    name = str(stage)
+    head, sep, tail = name.partition(".")
+    if sep and head.endswith("Pipeline"):
+        name = tail
+    if name.endswith("_ms"):
+        name = name[: -len("_ms")]
+    return name.replace("_", " ").replace(".", " ").title()
+
+
+def _is_printed_stage_gen_key(stage: str) -> bool:
+    """True for ``stage_0_gen_ms``, which duplicates ``--print-stage`` stage timing."""
+    name = str(stage)
+    prefix, suffix = "stage_", "_gen_ms"
+    if not name.startswith(prefix) or not name.endswith(suffix):
+        return False
+    return name[len(prefix) : -len(suffix)].isdigit()
+
+
+def print_stage_durations_metrics(
+    selected_percentiles: list[float],
+    metrics: MultiModalsBenchmarkMetrics,
+) -> None:
+    """Print profiler timings already stored on ``metrics``.
+
+    Mean and median always print, matching :func:`print_image_metrics`. P99 prints
+    only when 99 is in ``selected_percentiles``. ``stage_N_gen_ms`` is not here;
+    that clock stays on stage_gen_time / image / video generation.
+    """
+    mean = getattr(metrics, defs.STAGE_DURATIONS_MEAN, None) or {}
+    if not mean:
+        return
+    p50 = getattr(metrics, defs.STAGE_DURATIONS_P50, None) or {}
+    p99 = getattr(metrics, defs.STAGE_DURATIONS_P99, None) or {}
+    show_p99 = any(float(p) == 99.0 for p in selected_percentiles)
+    print("{s:{c}^{n}}".format(s=" Stage Durations ", n=50, c="-"))
+    for stage in mean:
+        label = _stage_duration_display_name(stage)
+        unit = " (ms)" if str(stage).endswith("_ms") else " (s)"
+        print("{:<40} {:<10.2f}".format(f"Mean {label}{unit}:", mean[stage]))
+        if stage in p50:
+            print("{:<40} {:<10.2f}".format(f"Median {label}{unit}:", p50[stage]))
+        if show_p99 and stage in p99:
+            print("{:<40} {:<10.2f}".format(f"P99 {label}{unit}:", p99[stage]))
+
+
 def print_image_metrics(selected_percentiles: list[float], metrics: MultiModalsBenchmarkMetrics):
     print("{s:{c}^{n}}".format(s=" Image Result ", n=50, c="="))
     print("{:<40} {:<10}".format("Total images generated:", getattr(metrics, defs.TOTAL_IMAGES)))
@@ -818,7 +912,7 @@ def calculate_metrics(
     request_rate,
     benchmark_duration,
     print_stage: bool = False,
-) -> tuple[BenchmarkMetrics, list[int]]:
+) -> tuple[MultiModalsBenchmarkMetrics, list[int]]:
     """Calculate the metrics for the benchmark.
 
     Args:
@@ -928,7 +1022,7 @@ def calculate_metrics(
             goodput_ttfts.append(outputs[i].ttft if ttft_measured else None)
             goodput_audio_ttfps.append(getattr(outputs[i], defs.AUDIO_TTFP, 0.0) if audio_ttfp_measured else None)
             audio_duration.append(getattr(outputs[i], defs.AUDIO_DURATION, 0.0))
-            audio_frames.append(getattr(outputs[i], defs.AUDIO_FRAMES, 0.0))
+            audio_frames.append(getattr(outputs[i], defs.AUDIO_FRAMES, 0))
             image_count = int(getattr(outputs[i], defs.IMAGE_COUNT, 0) or 0)
             total_images += image_count
             image_generation_time_ms = float(getattr(outputs[i], defs.IMAGE_GENERATION_TIME_MS, 0.0) or 0.0)
@@ -982,9 +1076,17 @@ def calculate_metrics(
                 good_completed += 1
 
     if completed == 0:
-        warnings.formatwarning = lambda msg, category, filename, lineno, line=None: (
-            f"{filename}:{lineno}: {category.__name__}: {msg}\n"
-        )
+
+        def _formatwarning(
+            message: Warning | str,
+            category: type[Warning],
+            filename: str,
+            lineno: int,
+            line: str | None = None,
+        ) -> str:
+            return f"{filename}:{lineno}: {category.__name__}: {message}\n"
+
+        warnings.formatwarning = _formatwarning
         warnings.warn(
             "All requests failed. This is likely due to a misconfiguration on the benchmark arguments.",
             stacklevel=2,
@@ -1074,6 +1176,7 @@ def calculate_metrics(
         isinstance(getattr(output, "duplex_session_metrics", None), dict) for output in outputs
     )
     missing_duplex_value = float("nan") if duplex_metrics_present else 0
+    stage_duration_summaries = aggregate_stage_durations(outputs)
     metrics = MultiModalsBenchmarkMetrics(
         completed=completed,
         failed=len(failed_outputs),
@@ -1167,6 +1270,9 @@ def calculate_metrics(
             defs.AUDIO_CONTINUITY_OK_RATE: (
                 (sum(audio_continuity_ok) / len(audio_continuity_ok)) if audio_continuity_ok else 1.0
             ),
+            defs.STAGE_DURATIONS_MEAN: stage_duration_summaries.get(defs.STAGE_DURATIONS_MEAN) or {},
+            defs.STAGE_DURATIONS_P50: stage_duration_summaries.get(defs.STAGE_DURATIONS_P50) or {},
+            defs.STAGE_DURATIONS_P99: stage_duration_summaries.get(defs.STAGE_DURATIONS_P99) or {},
         },
     )
     print_metrics(
diff --git a/vllm_omni/benchmarks/patch/patch.py b/vllm_omni/benchmarks/patch/patch.py
index 87fdb2daa1e..cfc609c3e34 100644
--- a/vllm_omni/benchmarks/patch/patch.py
+++ b/vllm_omni/benchmarks/patch/patch.py
@@ -19,6 +19,7 @@
 from datetime import datetime
 from pathlib import Path
 from typing import TYPE_CHECKING, Any, Literal
+from urllib.parse import urlparse
 
 import aiohttp
 import numpy as np
@@ -873,6 +874,9 @@ class MixRequestFuncOutput(RequestFuncOutput):
     tts_turn_pcm_bytes: list[bytes] | None = None
     #: Per-stage snapshot from orchestrator ``metrics["stage_metrics"]`` (merged across SSE chunks).
     stage_metrics: dict[str, dict] | None = None
+    #: Diffusion pipeline profiler timings from response ``stage_durations``
+    #: (e.g. diffuse / text_encoder.forward / vae.decode), when present.
+    stage_durations: dict[str, float] | None = None
     stage_id: int | None = None
     final_output_type: str | None = None
     duplex_request_metrics: list[dict[str, object]] | None = None
@@ -883,6 +887,7 @@ class MixRequestFuncOutput(RequestFuncOutput):
     "negative_prompt",
     "num_inference_steps",
     "guidance_scale",
+    "guidance_scale_2",
     "strength",
     "true_cfg_scale",
     "seed",
@@ -902,15 +907,51 @@ def _guess_mime_type(path: str) -> str:
     return mime or "application/octet-stream"
 
 
-def _iter_image_edit_inputs(value: Any) -> Iterable[Any]:
-    """Yield image references from benchmark multimodal content."""
+_IMAGE_REFERENCE_SUFFIXES = frozenset({".png", ".jpg", ".jpeg", ".webp", ".gif", ".bmp", ".heic", ".heif"})
+_VIDEO_REFERENCE_SUFFIXES = frozenset({".mp4", ".mov", ".webm", ".mkv", ".m4v"})
+
+
+def _string_reference_kind(reference: str) -> str | None:
+    """Classify a bare reference string as image, video, or file.
+
+    ``data:image`` / ``data:video`` carry their type in the URL. Bare http(s)
+    URLs use the path extension. An existing local path is a file upload, not
+    an image or a video URL.
+    """
+    if reference.startswith("data:image"):
+        return "image"
+    if reference.startswith("data:video"):
+        return "video"
+    if reference.startswith(("http://", "https://")):
+        suffix = Path(urlparse(reference).path).suffix.lower()
+        if suffix in _VIDEO_REFERENCE_SUFFIXES:
+            return "video"
+        if suffix in _IMAGE_REFERENCE_SUFFIXES:
+            return "image"
+        return None
+    local_path = reference.removeprefix("file://")
+    if local_path and os.path.exists(local_path):
+        return "file"
+    return None
+
+
+def _iter_image_reference_inputs(value: Any) -> Iterable[Any]:
+    """Yield image references from benchmark multimodal content.
+
+    ``random-mm`` image buckets arrive as OpenAI chat parts
+    ``{"type": "image_url", "image_url": {"url": ...}}``. Yield
+    ``{"image_url": url}`` so the form helper keeps an explicit image type.
+    Bare video strings are left for ``_iter_video_reference_inputs``.
+    """
     if value is None:
         return
     if isinstance(value, list):
         for item in value:
-            yield from _iter_image_edit_inputs(item)
+            yield from _iter_image_reference_inputs(item)
         return
     if not isinstance(value, dict):
+        if isinstance(value, str) and _string_reference_kind(value) == "video":
+            return
         yield value
         return
 
@@ -919,15 +960,49 @@ def _iter_image_edit_inputs(value: Any) -> Iterable[Any]:
         image_url = value.get("image_url")
         if isinstance(image_url, dict):
             url = image_url.get("url")
-            if url:
-                yield url
-        elif image_url:
-            yield image_url
+            if isinstance(url, str) and url:
+                yield {"image_url": url}
+        elif isinstance(image_url, str) and image_url:
+            yield {"image_url": image_url}
         return
 
     for key in ("image", "images"):
         if key in value:
-            yield from _iter_image_edit_inputs(value[key])
+            yield from _iter_image_reference_inputs(value[key])
+
+
+def _iter_video_reference_inputs(value: Any) -> Iterable[dict[str, str]]:
+    """Yield structured video references from benchmark multimodal content.
+
+    ``random-mm`` video buckets arrive as OpenAI chat parts
+    ``{"type": "video_url", "video_url": {"url": ...}}``, or as a bare
+    ``data:video`` / video http(s) string. Yield ``{"video_url": url}`` so the
+    form helper keeps the video branch.
+    """
+    if value is None:
+        return
+    if isinstance(value, list):
+        for item in value:
+            yield from _iter_video_reference_inputs(item)
+        return
+    if not isinstance(value, dict):
+        if isinstance(value, str) and _string_reference_kind(value) == "video":
+            yield {"video_url": value}
+        return
+
+    if value.get("type") == "video_url":
+        video_url = value.get("video_url")
+        if isinstance(video_url, dict):
+            url = video_url.get("url")
+            if isinstance(url, str) and url:
+                yield {"video_url": url}
+        elif isinstance(video_url, str) and video_url:
+            yield {"video_url": video_url}
+        return
+
+    for key in ("video", "videos"):
+        if key in value:
+            yield from _iter_video_reference_inputs(value[key])
 
 
 def _add_image_edit_input_to_form(form: aiohttp.FormData, image_input: Any) -> None:
@@ -940,6 +1015,12 @@ def _add_image_edit_input_to_form(form: aiohttp.FormData, image_input: Any) -> N
         )
         return
 
+    if isinstance(image_input, Mapping) and _is_structured_image_reference(image_input):
+        image_url = image_input.get("image_url")
+        if isinstance(image_url, str) and image_url:
+            _add_image_edit_input_to_form(form, image_url)
+            return
+
     if isinstance(image_input, str):
         if image_input.startswith(("data:image", "http://", "https://")):
             form.add_field("url", image_input)
@@ -1137,6 +1218,68 @@ def _update_output_peak_memory_from_payload(output: MixRequestFuncOutput, data:
         output.peak_memory_mb = peak_memory_mb
 
 
+def _coerce_stage_durations_dict(raw: object) -> dict[str, float] | None:
+    """Normalize a stage_durations mapping to ``dict[str, float]``."""
+    if not isinstance(raw, dict) or not raw:
+        return None
+    coerced: dict[str, float] = {}
+    for key, value in raw.items():
+        if isinstance(value, bool) or not isinstance(value, (int, float)) or not np.isfinite(value):
+            continue
+        coerced[str(key)] = float(value)
+    return coerced or None
+
+
+def _extract_stage_durations_from_payload(data: Mapping[str, object]) -> dict[str, float] | None:
+    """Pull pipeline profiler timings from video/image/chat response shapes."""
+    found = _coerce_stage_durations_dict(data.get("stage_durations"))
+    if found:
+        return found
+
+    metrics = data.get("metrics")
+    if isinstance(metrics, dict):
+        found = _coerce_stage_durations_dict(metrics.get("stage_durations"))
+        if found:
+            return found
+
+    response_data = data.get("data")
+    if isinstance(response_data, list):
+        for item in response_data:
+            if isinstance(item, dict):
+                found = _coerce_stage_durations_dict(item.get("stage_durations"))
+                if found:
+                    return found
+
+    choices = data.get("choices")
+    if isinstance(choices, list):
+        for choice in choices:
+            if not isinstance(choice, dict):
+                continue
+            for message_key in ("message", "delta"):
+                message = choice.get(message_key)
+                if not isinstance(message, dict):
+                    continue
+                content = message.get("content")
+                if isinstance(content, list):
+                    for item in content:
+                        if isinstance(item, dict):
+                            found = _coerce_stage_durations_dict(item.get("stage_durations"))
+                            if found:
+                                return found
+                elif isinstance(content, dict):
+                    found = _coerce_stage_durations_dict(content.get("stage_durations"))
+                    if found:
+                        return found
+    return None
+
+
+def _update_output_stage_durations_from_payload(output: MixRequestFuncOutput, data: Mapping[str, object]) -> None:
+    """Persist the full profiler ``stage_durations`` map when the response has one."""
+    found = _extract_stage_durations_from_payload(data)
+    if found:
+        output.stage_durations = found
+
+
 def _image_metrics_from_stage_metrics(metrics: object) -> tuple[int, float, int, float]:
     if not isinstance(metrics, dict):
         return 0, 0.0, 0, 0.0
@@ -1209,6 +1352,7 @@ def _apply_image_metrics_from_payload(output: MixRequestFuncOutput, data: Mappin
     """Populate image benchmark fields from an OpenAI-compatible image payload."""
     _update_output_stage_metrics_from_payload(output, data, update_output_tokens=False)
     _update_output_peak_memory_from_payload(output, data)
+    _update_output_stage_durations_from_payload(output, data)
 
     payload_image_count = 0
     response_data = data.get("data")
@@ -1321,15 +1465,129 @@ def _video_frames_from_payload(data: Mapping[str, object], request_body: Mapping
 
 
 def _is_structured_image_reference(reference: Mapping[str, object]) -> bool:
-    """True for API image_reference objects ({image_url}/{file_id})."""
+    """True for API image_reference objects ({"image_url": "..."})."""
     image_url = reference.get("image_url")
-    file_id = reference.get("file_id")
-    has_url = isinstance(image_url, str) and bool(image_url)
-    has_file_id = isinstance(file_id, str) and bool(file_id)
-    return has_url or has_file_id
+    return isinstance(image_url, str) and bool(image_url)
+
+
+def _is_structured_video_reference(reference: Mapping[str, object]) -> bool:
+    """True for API video_reference objects ({"video_url": "..."})."""
+    video_url = reference.get("video_url")
+    return isinstance(video_url, str) and bool(video_url)
+
+
+_VIDEO_REFERENCE_JSON_MAX_BYTES = 1024 * 1024
+
+
+def _data_video_json_exceeds_text_limit(video_url: str) -> bool:
+    """True when a data:video URL would exceed the ~1MB multipart text-part limit."""
+    if not video_url.startswith("data:video"):
+        return False
+    encoded = json.dumps({"video_url": video_url}).encode("utf-8")
+    return len(encoded) > _VIDEO_REFERENCE_JSON_MAX_BYTES
+
+
+def _add_data_video_upload(form: aiohttp.FormData, video_url: str) -> bool:
+    """Upload one inline video as ``input_references`` instead of a JSON text part."""
+    header, _, payload = video_url.partition(",")
+    if not payload:
+        raise ValueError(f"Unsupported video data URL: {video_url[:64]!r}")
+    try:
+        video_bytes = base64.b64decode(payload)
+    except (ValueError, TypeError) as exc:
+        raise ValueError("video data URL is not valid base64") from exc
+    mime = header[len("data:") :].split(";", 1)[0] or "video/mp4"
+    suffix = ".mp4" if mime.endswith("mp4") else ".bin"
+    form.add_field(
+        "input_references",
+        video_bytes,
+        filename=f"benchmark-reference{suffix}",
+        content_type=mime,
+    )
+    return True
 
 
-def _add_video_reference_to_form(form: aiohttp.FormData, reference: object) -> bool:
+def _file_bytes_as_data_url(raw: bytes, mime: str) -> str:
+    encoded = base64.b64encode(raw).decode("ascii")
+    return f"data:{mime};base64,{encoded}"
+
+
+def _image_reference_json_value(reference: object) -> object:
+    """Turn an image file into a JSON ``image_reference`` when it must share a form.
+
+    ``input_reference`` cannot be combined with ``video_reference``. Image URLs
+    stay URLs. Upload bytes and local image files become ``data:image`` URLs.
+    """
+    if isinstance(reference, Mapping) and "bytes" in reference and not _is_structured_image_reference(reference):
+        raw = reference["bytes"]
+        if not isinstance(raw, (bytes, bytearray)):
+            raise ValueError(f"image reference bytes must be bytes (got {type(raw).__name__}).")
+        content_type = reference.get("content_type", "image/png")
+        if not isinstance(content_type, str) or not content_type.startswith("image/"):
+            content_type = "image/png"
+        return {"image_url": _file_bytes_as_data_url(bytes(raw), content_type.split(";", 1)[0])}
+    if isinstance(reference, str):
+        kind = _string_reference_kind(reference)
+        if kind == "image":
+            return {"image_url": reference}
+        if kind == "file":
+            local_path = reference.removeprefix("file://")
+            mime = _guess_mime_type(local_path)
+            if not mime.startswith("image/"):
+                mime = "image/png"
+            with open(local_path, "rb") as handle:
+                return {"image_url": _file_bytes_as_data_url(handle.read(), mime)}
+    return reference
+
+
+def _video_reference_json_value(reference: object) -> object:
+    """Turn a video file into a JSON ``video_reference`` when it must share a form.
+
+    ``input_reference`` cannot be combined with ``image_reference``. Video URLs
+    stay URLs. Local video files become ``data:video`` URLs.
+    """
+    if isinstance(reference, Mapping) and "bytes" in reference and not _is_structured_video_reference(reference):
+        raw = reference["bytes"]
+        if not isinstance(raw, (bytes, bytearray)):
+            raise ValueError(f"video reference bytes must be bytes (got {type(raw).__name__}).")
+        content_type = reference.get("content_type", "video/mp4")
+        if not isinstance(content_type, str) or not content_type.startswith("video/"):
+            content_type = "video/mp4"
+        return {"video_url": _file_bytes_as_data_url(bytes(raw), content_type.split(";", 1)[0])}
+    if isinstance(reference, str):
+        kind = _string_reference_kind(reference)
+        if kind == "video":
+            return {"video_url": reference}
+        if kind == "file":
+            local_path = reference.removeprefix("file://")
+            mime = _guess_mime_type(local_path)
+            if not mime.startswith("video/"):
+                mime = "video/mp4"
+            with open(local_path, "rb") as handle:
+                return {"video_url": _file_bytes_as_data_url(handle.read(), mime)}
+    return reference
+
+
+def _add_video_reference_to_form(
+    form: aiohttp.FormData,
+    reference: object,
+    *,
+    upload_inline_video: bool = True,
+) -> bool:
+    """Encode one reference: image URL, video URL, or file upload.
+
+    Image URLs use ``image_reference``. Video URLs use ``video_reference``.
+    A lone ``data:video`` whose JSON text exceeds 1MB is uploaded as
+    ``input_references``. Local paths and raw bytes use ``input_reference``.
+    ``upload_inline_video`` must be false when an image is on the same form:
+    ``input_references`` cannot be combined with ``image_reference``.
+    """
+    candidates = reference if isinstance(reference, list) else [reference]
+    for item in candidates:
+        if isinstance(item, Mapping):
+            file_id = item.get("file_id")
+            if isinstance(file_id, str) and file_id:
+                raise ValueError("file_id is not supported yet")
     if isinstance(reference, dict) and "bytes" in reference:
         form.add_field(
             "input_reference",
@@ -1343,23 +1601,39 @@ def _add_video_reference_to_form(form: aiohttp.FormData, reference: object) -> b
         form.add_field("image_reference", json.dumps(dict(reference)))
         return True
 
+    if isinstance(reference, Mapping) and _is_structured_video_reference(reference):
+        video_url = reference.get("video_url")
+        if upload_inline_video and isinstance(video_url, str) and _data_video_json_exceeds_text_limit(video_url):
+            return _add_data_video_upload(form, video_url)
+        form.add_field("video_reference", json.dumps(dict(reference)))
+        return True
+
     if isinstance(reference, list):
         if reference and all(isinstance(item, Mapping) and _is_structured_image_reference(item) for item in reference):
             form.add_field("image_reference", json.dumps([dict(item) for item in reference]))
             return True
+        if reference and all(isinstance(item, Mapping) and _is_structured_video_reference(item) for item in reference):
+            form.add_field("video_reference", json.dumps([dict(item) for item in reference]))
+            return True
         raise ValueError(
-            "Unsupported image_reference list; expected non-empty list of "
-            '{"image_url": "..."} and/or {"file_id": "..."} objects.'
+            "Unsupported reference list; expected non-empty list of "
+            '{"image_url": "..."} or {"video_url": "..."} objects.'
         )
 
     if isinstance(reference, str):
-        if reference.startswith(("data:image", "http://", "https://")):
+        kind = _string_reference_kind(reference)
+        if kind == "image":
             form.add_field("image_reference", json.dumps({"image_url": reference}))
             return True
-        local_path = reference.removeprefix("file://")
-        if os.path.exists(local_path):
-            with open(local_path, "rb") as f:
-                reference_bytes = f.read()
+        if kind == "video":
+            if upload_inline_video and _data_video_json_exceeds_text_limit(reference):
+                return _add_data_video_upload(form, reference)
+            form.add_field("video_reference", json.dumps({"video_url": reference}))
+            return True
+        if kind == "file":
+            local_path = reference.removeprefix("file://")
+            with open(local_path, "rb") as handle:
+                reference_bytes = handle.read()
             form.add_field(
                 "input_reference",
                 reference_bytes,
@@ -1367,15 +1641,64 @@ def _add_video_reference_to_form(form: aiohttp.FormData, reference: object) -> b
                 content_type=_guess_mime_type(local_path),
             )
             return True
-        raise ValueError(f"Unsupported image_reference path or URL: {reference!r}")
+        if reference.startswith(("http://", "https://")):
+            raise ValueError(
+                "Bare http(s) reference needs an image or video extension "
+                f"({', '.join(sorted(_IMAGE_REFERENCE_SUFFIXES | _VIDEO_REFERENCE_SUFFIXES))}); "
+                f"got {reference!r}."
+            )
+        raise ValueError(f"Unsupported reference path or URL: {reference!r}")
 
     raise ValueError(
-        "Unsupported image_reference; expected upload bytes, local path/URL string, "
-        'or {"image_url": "..."} / {"file_id": "..."} object '
+        "Unsupported reference; expected image URL, video URL, upload bytes, or a local file "
         f"(got {type(reference).__name__})."
     )
 
 
+def _add_combined_video_form_references(
+    form: aiohttp.FormData,
+    multi_modal_content: Any,
+    extra_body: Mapping[str, Any] | None = None,
+) -> None:
+    """Serialize image and video refs using a server-accepted field pair.
+
+    Alone, each reference uses its own field: image URL → ``image_reference``,
+    video URL → ``video_reference``, file → ``input_reference``. A lone inline
+    video whose JSON text exceeds 1MB is uploaded as ``input_references``.
+    Together, that upload cannot be combined with ``image_reference``, so both
+    sides stay on the JSON fields. A file paired with the other media is rewritten
+    as a data URL of the matching type.
+    """
+    extra_body = extra_body or {}
+    image_refs = list(_iter_image_reference_inputs(multi_modal_content))
+    video_refs = list(_iter_video_reference_inputs(multi_modal_content))
+    if not image_refs and extra_body.get("image_reference") is not None:
+        image_refs = [extra_body["image_reference"]]
+    if not video_refs and extra_body.get("video_reference") is not None:
+        video_refs = [extra_body["video_reference"]]
+
+    if image_refs and video_refs:
+        for raw in (image_refs[0], video_refs[0]):
+            candidates = raw if isinstance(raw, list) else [raw]
+            for item in candidates:
+                if isinstance(item, Mapping):
+                    file_id = item.get("file_id")
+                    if isinstance(file_id, str) and file_id:
+                        raise ValueError("file_id is not supported yet")
+        _add_video_reference_to_form(form, _image_reference_json_value(image_refs[0]))
+        _add_video_reference_to_form(
+            form,
+            _video_reference_json_value(video_refs[0]),
+            upload_inline_video=False,
+        )
+        return
+
+    if image_refs:
+        _add_video_reference_to_form(form, image_refs[0])
+    if video_refs:
+        _add_video_reference_to_form(form, video_refs[0])
+
+
 def _add_video_extra_body_to_form(
     form: aiohttp.FormData,
     extra_body: Mapping[str, object],
@@ -1398,9 +1721,11 @@ def _add_video_extra_body_to_form(
         "height",
         "poll_interval_s",
         "poll_timeout_s",
-        # Handled only by _add_video_reference_to_form (upload / JSON image_url).
+        # Handled only by _add_video_reference_to_form (upload / JSON image_url / video_url).
         "image_reference",
+        "video_reference",
         "input_reference",
+        "input_references",
         *_VIDEO_FORM_FIELDS,
     }
     for key, value in extra_body.items():
@@ -1421,8 +1746,9 @@ def _apply_video_metrics_from_payload(
     output.video_frames = _video_frames_from_payload(data, request_body)
     _update_output_stage_metrics_from_payload(output, data, update_output_tokens=False)
     _update_output_peak_memory_from_payload(output, data)
+    _update_output_stage_durations_from_payload(output, data)
 
-    stage_durations = data.get("stage_durations")
+    stage_durations = output.stage_durations if output.stage_durations is not None else data.get("stage_durations")
     stage_gen_ms = _video_generation_ms_from_stage_durations(stage_durations)
     if stage_gen_ms <= 0:
         inference_time_s = coerce_positive_float_scalar(data.get("inference_time_s"))
@@ -1534,6 +1860,7 @@ async def async_request_openai_chat_omni_completions(
         output.image_pixels = 0
         output.denoise_step_latency_ms = 0.0
         output.peak_memory_mb = 0.0
+        output.stage_durations = None
         completion_tokens_seen = 0
         try:
             async with session.post(url=api_url, json=payload, headers=headers) as response:
@@ -1564,6 +1891,7 @@ async def async_request_openai_chat_omni_completions(
                                 data = json.loads(chunk)
                                 _update_output_stage_metrics_from_payload(output, data)
                                 _update_output_peak_memory_from_payload(output, data)
+                                _update_output_stage_durations_from_payload(output, data)
                                 usage = data.get("usage")
                                 completion_tokens = None
                                 if isinstance(usage, dict):
@@ -1926,15 +2254,7 @@ async def async_request_openai_videos_omni(
         form.add_field("size", str(request_body["size"]))
     _add_video_extra_body_to_form(form, extra_body, request_body)
 
-    reference_added = False
-    for reference in _iter_image_edit_inputs(request_func_input.multi_modal_content):
-        if _add_video_reference_to_form(form, reference):
-            reference_added = True
-            break
-    if not reference_added:
-        image_reference = extra_body.get("image_reference")
-        if image_reference is not None:
-            _add_video_reference_to_form(form, image_reference)
+    _add_combined_video_form_references(form, request_func_input.multi_modal_content, extra_body)
 
     headers = {
         "Authorization": f"Bearer {os.environ.get('OPENAI_API_KEY')}",
@@ -2044,7 +2364,7 @@ async def async_request_openai_image_edits_omni(
     _add_image_edit_extra_body_to_form(form, extra_body)
 
     try:
-        image_inputs = list(_iter_image_edit_inputs(request_func_input.multi_modal_content))
+        image_inputs = list(_iter_image_reference_inputs(request_func_input.multi_modal_content))
         if not image_inputs:
             raise ValueError(
                 "openai-image-edits-omni requires image multimodal content. "
@@ -2102,6 +2422,7 @@ async def async_request_openai_image_edits_omni(
                             update_output_tokens=(data.get("type") == "ar_delta"),
                         )
                         _update_output_peak_memory_from_payload(output, data)
+                        _update_output_stage_durations_from_payload(output, data)
 
                         chunk_type = data.get("type")
                         if chunk_type == "ar_delta":
@@ -3201,6 +3522,9 @@ def measured_ttft(output: RequestFuncOutput) -> float | None:
             defs.MEAN_PEAK_MEMORY_MB: getattr(mm_metrics, defs.MEAN_PEAK_MEMORY_MB),
             defs.MEDIAN_PEAK_MEMORY_MB: getattr(mm_metrics, defs.MEDIAN_PEAK_MEMORY_MB),
             defs.PERCENTILES_PEAK_MEMORY_MB: getattr(mm_metrics, defs.PERCENTILES_PEAK_MEMORY_MB),
+            defs.STAGE_DURATIONS_MEAN: getattr(mm_metrics, defs.STAGE_DURATIONS_MEAN) or {},
+            defs.STAGE_DURATIONS_P50: getattr(mm_metrics, defs.STAGE_DURATIONS_P50) or {},
+            defs.STAGE_DURATIONS_P99: getattr(mm_metrics, defs.STAGE_DURATIONS_P99) or {},
             "input_lens": [output.prompt_len for output in outputs],
             "start_times": [output.start_time for output in outputs],
             "output_lens": actual_output_lens,
diff --git a/vllm_omni/diffusion/models/cosmos3/pipeline_cosmos3.py b/vllm_omni/diffusion/models/cosmos3/pipeline_cosmos3.py
index 04f25ac58d9..95ba4767fe9 100644
--- a/vllm_omni/diffusion/models/cosmos3/pipeline_cosmos3.py
+++ b/vllm_omni/diffusion/models/cosmos3/pipeline_cosmos3.py
@@ -1730,7 +1730,10 @@ def _forward_robolab_policy(
                 },
             },
         }
-        return DiffusionOutput(output=action_output)
+        return DiffusionOutput(
+            output=action_output,
+            stage_durations=self.stage_durations if hasattr(self, "stage_durations") else None,
+        )
 
     @staticmethod
     def _truthy(value) -> bool:
@@ -3610,6 +3613,7 @@ def _forward_transfer(
                     "payload": {"video": output_video},
                     "metadata": {"video": {"fps": frame_rate}},
                 },
+                stage_durations=self.stage_durations if hasattr(self, "stage_durations") else None,
             )
 
         full_output = torch.cat(output_chunks, dim=2)[:, :, :total_frames]
@@ -3636,6 +3640,7 @@ def _forward_transfer(
                     "video": {"fps": frame_rate},
                 },
             },
+            stage_durations=self.stage_durations if hasattr(self, "stage_durations") else None,
         )
 
     # -- Forward (main generation entry point) -------------------------------
@@ -4072,7 +4077,10 @@ def _run_diffusion(start_latents):
             if _is_rank_zero():
                 logger.info("Sound tokenizer decoded in %.2fs", time.time() - sound_decode_start)
                 logger.info("Total pipeline time: %.2fs", time.time() - pipeline_start)
-            return DiffusionOutput(output={"video": video, "audio": audio, "audio_sample_rate": sound_sample_rate})
+            return DiffusionOutput(
+                output={"video": video, "audio": audio, "audio_sample_rate": sound_sample_rate},
+                stage_durations=self.stage_durations if hasattr(self, "stage_durations") else None,
+            )
 
         if action_enabled:
             if action_latents is None or raw_action_dim is None or domain_id is None:
@@ -4092,6 +4100,10 @@ def _run_diffusion(start_latents):
                         },
                     },
                 },
+                stage_durations=self.stage_durations if hasattr(self, "stage_durations") else None,
             )
 
-        return DiffusionOutput(output={"image": video} if is_t2i else {"video": video})
+        return DiffusionOutput(
+            output={"image": video} if is_t2i else {"video": video},
+            stage_durations=self.stage_durations if hasattr(self, "stage_durations") else None,
+        )
diff --git a/vllm_omni/diffusion/models/hunyuan_video/pipeline_hunyuan_video_1_5.py b/vllm_omni/diffusion/models/hunyuan_video/pipeline_hunyuan_video_1_5.py
index 942e0b28435..5f0ecb452e1 100644
--- a/vllm_omni/diffusion/models/hunyuan_video/pipeline_hunyuan_video_1_5.py
+++ b/vllm_omni/diffusion/models/hunyuan_video/pipeline_hunyuan_video_1_5.py
@@ -1,5 +1,5 @@
 # SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM-Omni project
 
 from __future__ import annotations
 
@@ -548,7 +548,10 @@ def forward(
             latents = latents.to(self.vae.dtype) / self.vae.config.scaling_factor
             output = self.vae.decode(latents, return_dict=False)[0]
 
-        return DiffusionOutput(output=output)
+        return DiffusionOutput(
+            output=output,
+            stage_durations=self.stage_durations if hasattr(self, "stage_durations") else None,
+        )
 
     def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
         loader = AutoWeightsLoader(self)
diff --git a/vllm_omni/diffusion/models/hunyuan_video/pipeline_hunyuan_video_1_5_i2v.py b/vllm_omni/diffusion/models/hunyuan_video/pipeline_hunyuan_video_1_5_i2v.py
index 1281a9e39d6..f01d45a4ec9 100644
--- a/vllm_omni/diffusion/models/hunyuan_video/pipeline_hunyuan_video_1_5_i2v.py
+++ b/vllm_omni/diffusion/models/hunyuan_video/pipeline_hunyuan_video_1_5_i2v.py
@@ -1,5 +1,5 @@
 # SPDX-License-Identifier: Apache-2.0
-# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM-Omni project
 
 from __future__ import annotations
 
@@ -643,7 +643,10 @@ def forward(
             latents = latents.to(self.vae.dtype) / self.vae.config.scaling_factor
             output = self.vae.decode(latents, return_dict=False)[0]
 
-        return DiffusionOutput(output=output)
+        return DiffusionOutput(
+            output=output,
+            stage_durations=self.stage_durations if hasattr(self, "stage_durations") else None,
+        )
 
     def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
         loader = AutoWeightsLoader(self)
diff --git a/vllm_omni/metrics/definitions.py b/vllm_omni/metrics/definitions.py
index a09fbcc2ff2..59b575e50fc 100644
--- a/vllm_omni/metrics/definitions.py
+++ b/vllm_omni/metrics/definitions.py
@@ -99,6 +99,10 @@
 STD_PEAK_MEMORY_MB = f"std_{PEAK_MEMORY_MB}"
 PERCENTILES_PEAK_MEMORY_MB = f"percentiles_{PEAK_MEMORY_MB}"
 
+STAGE_DURATIONS_MEAN = "stage_durations_mean"
+STAGE_DURATIONS_P50 = "stage_durations_p50"
+STAGE_DURATIONS_P99 = "stage_durations_p99"
+
 # Stage snapshot / StageBenchmarkMetrics field names.
 TOTAL_OUTPUT = "total_output"
 TTFTS = "ttfts"