[Nightly][Refactor]Migrate nightly single-node model tests from .py to .yaml (#6503)

### What this PR does / why we need it? This PR refactors the nightly single-node model test by migrating test configurations from Python scripts to a more maintainable `YAML-based` format. | Original PR | Python (`.py`) | YAML (`.yaml`) | | :--- | :--- | :--- | | [#3568](https://github.com/vllm-project/vllm-ascend/pull/3568) | `test_deepseek_r1_0528_w8a8_eplb.py` | `DeepSeek-R1-0528-W8A8.yaml` | | [#3631](https://github.com/vllm-project/vllm-ascend/pull/3631) | `test_deepseek_r1_0528_w8a8.py` | `DeepSeek-R1-0528-W8A8.yaml` | | [#5874](https://github.com/vllm-project/vllm-ascend/pull/5874) | `test_deepseek_r1_w8a8_hbm.py` | `DeepSeek-R1-W8A8-HBM.yaml` | | [#3908](https://github.com/vllm-project/vllm-ascend/pull/3908) | `test_deepseek_v3_2_w8a8.py` | `DeepSeek-V3.2-W8A8.yaml` | | [#5682](https://github.com/vllm-project/vllm-ascend/pull/5682) | `test_kimi_k2_thinking.py` | `Kimi-K2-Thinking.yaml` | | [#4111](https://github.com/vllm-project/vllm-ascend/pull/4111) | `test_mtpx_deepseek_r1_0528_w8a8.py` | `MTPX-DeepSeek-R1-0528-W8A8.yaml` | | [#3733](https://github.com/vllm-project/vllm-ascend/pull/3733) | `test_prefix_cache_deepseek_r1_0528_w8a8.py` | `Prefix-Cache-DeepSeek-R1-0528-W8A8.yaml` | | [#6543](https://github.com/vllm-project/vllm-ascend/pull/6543) | `test_qwen3_235b_w8a8.py` | `Qwen3-235B-A22B-W8A8.yaml` | | [#6543](https://github.com/vllm-project/vllm-ascend/pull/6543) | `test_qwen3_235b_a22b_w8a8_eplb.py` | `Qwen3-235B-A22B-W8A8.yaml` | | [#3973](https://github.com/vllm-project/vllm-ascend/pull/3973) | `test_qwen3_30b_w8a8.py` | `Qwen3-30B-A3B-W8A8.yaml` | | [#3541](https://github.com/vllm-project/vllm-ascend/pull/3541) | `test_qwen3_32b_int8.py` | `Qwen3-32B-Int8.yaml` | | [#3757](https://github.com/vllm-project/vllm-ascend/pull/3757) | `test_qwq_32b.py` | `QwQ-32B.yaml` | | [#5616](https://github.com/vllm-project/vllm-ascend/pull/5616) | `test_qwen3_next_w8a8.py` | `Qwen3-Next-80B-A3B-Instruct-W8A8.yaml` | | [#3541](https://github.com/vllm-project/vllm-ascend/pull/3541) | `test_qwen2_5_vl_7b.py` | `Qwen2.5-VL-7B-Instruct.yaml` | | [#5301](https://github.com/vllm-project/vllm-ascend/pull/5301) | `test_qwen2_5_vl_7b_epd.py` | `Qwen2.5-VL-7B-Instruct-EPD.yaml` | | [#3707](https://github.com/vllm-project/vllm-ascend/pull/3707) | `test_qwen2_5_vl_32b.py` | `Qwen2.5-VL-32B-Instruct.yaml` | | [#3676](https://github.com/vllm-project/vllm-ascend/pull/3676) | `test_qwen3_32b_int8_a3_feature_stack3.py` | `Qwen3-32B-Int8-A3-Feature-Stack3.yaml` | | [#3709](https://github.com/vllm-project/vllm-ascend/pull/3709) | `test_prefix_cache_qwen3_32b_int8.py` | `Prefix-Cache-Qwen3-32B-Int8.yaml` | | [#5395](https://github.com/vllm-project/vllm-ascend/pull/5395) | `test_qwen3_next.py` | `Qwen3-Next-80B-A3B-Instruct-A2.yaml` | | [#3474](https://github.com/vllm-project/vllm-ascend/pull/3474) | `test_qwen3_32b.py` | `Qwen3-32B.yaml` | | [#3541](https://github.com/vllm-project/vllm-ascend/pull/3541) | `test_qwen3_32b_int8.py` | `Qwen3-32B-Int8-A2.yaml` | ### Does this PR introduce _any_ user-facing change? ### How was this patch tested? - vLLM version: v0.15.0 - vLLM main: https://github.com/vllm-project/vllm/commit/v0.15.0 --------- Signed-off-by: MrZ20 <2609716663@qq.com>
2026-03-03 20:13:43 +08:00
parent a0a904a3d4
commit 859f2c25b9
51 changed files with 2265 additions and 2336 deletions
--- a/tests/e2e/nightly/single_node/models/configs/DeepSeek-R1-0528-W8A8.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/DeepSeek-R1-0528-W8A8.yaml
@@ -0,0 +1,94 @@
+# ==========================================
+# Shared Configurations
+# ==========================================
+
+_envs: &envs
+  OMP_NUM_THREADS: "10"
+  OMP_PROC_BIND: "false"
+  HCCL_BUFFSIZE: "1024"
+  PYTORCH_NPU_ALLOC_CONF: "expandable_segments:True"
+  SERVER_PORT: "DEFAULT_PORT"
+
+_server_cmd: &server_cmd
+  - "--quantization"
+  - "ascend"
+  - "--data-parallel-size"
+  - "2"
+  - "--tensor-parallel-size"
+  - "8"
+  - "--enable-expert-parallel"
+  - "--port"
+  - "$SERVER_PORT"
+  - "--seed"
+  - "1024"
+  - "--max-model-len"
+  - "36864"
+  - "--max-num-batched-tokens"
+  - "4096"
+  - "--max-num-seqs"
+  - "16"
+  - "--trust-remote-code"
+  - "--gpu-memory-utilization"
+  - "0.9"
+  - "--speculative-config"
+  - '{"num_speculative_tokens": 1, "method": "mtp"}'
+  - "--additional-config"
+  - '{"enable_weight_nz_layout": true}'
+
+_benchmarks_acc: &benchmarks_acc
+  acc:
+    case_type: accuracy
+    dataset_path: vllm-ascend/gsm8k-lite
+    request_conf: vllm_api_general_chat
+    dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_chat_prompt
+    max_out_len: 32768
+    batch_size: 32
+    baseline: 95
+    threshold: 5
+
+_benchmarks_perf: &benchmarks_perf
+  perf:
+    case_type: performance
+    dataset_path: vllm-ascend/GSM8K-in3500-bs400
+    request_conf: vllm_api_stream_chat
+    dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+    num_prompts: 400
+    max_out_len: 1500
+    batch_size: 1000
+    baseline: 1
+    threshold: 0.97
+
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "DeepSeek-R1-0528-W8A8-single"
+    model: "vllm-ascend/DeepSeek-R1-0528-W8A8"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--enforce-eager"
+    benchmarks:
+
+  - name: "DeepSeek-R1-0528-W8A8-aclgraph"
+    model: "vllm-ascend/DeepSeek-R1-0528-W8A8"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    benchmarks:
+      <<: *benchmarks_acc
+      <<: *benchmarks_perf
+
+  - name: "DeepSeek-R1-0528-W8A8-EPLB"
+    model: "vllm-ascend/DeepSeek-R1-0528-W8A8"
+    envs:
+      <<: *envs
+      DYNAMIC_EPLB: "true"
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--additional-config"
+      - '{"enable_weight_nz_layout": true, "eplb_config": {"dynamic_eplb": "true", "expert_heat_collection_interval": 1000, "algorithm_execution_interval": 50, "eplb_policy_type": 3}}'
+    benchmarks:
+      <<: *benchmarks_acc
--- a/tests/e2e/nightly/single_node/models/configs/DeepSeek-R1-W8A8-HBM.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/DeepSeek-R1-W8A8-HBM.yaml
@@ -0,0 +1,42 @@
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "DeepSeek-R1-W8A8-HBM-single"
+    model: "vllm-ascend/DeepSeek-R1-W8A8"
+    envs:
+      HCCL_BUFFSIZE: "1024"
+      SERVER_PORT: "DEFAULT_PORT"
+    server_cmd:
+      - "--quantization"
+      - "ascend"
+      - "--port"
+      - "$SERVER_PORT"
+      - "--data-parallel-size"
+      - "8"
+      - "--data-parallel-size-local"
+      - "8"
+      - "--data-parallel-rpc-port"
+      - "13389"
+      - "--tensor-parallel-size"
+      - "2"
+      - "--enable-expert-parallel"
+      - "--seed"
+      - "1024"
+      - "--max-num-seqs"
+      - "32"
+      - "--max-model-len"
+      - "6000"
+      - "--max-num-batched-tokens"
+      - "6000"
+      - "--trust-remote-code"
+      - "--gpu-memory-utilization"
+      - "0.92"
+      - "--no-enable-prefix-caching"
+      - "--reasoning-parser"
+      - "deepseek_r1"
+      - "--enforce-eager"
+      - "--additional-config"
+      - '{"ascend_scheduler_config": {"enabled": false}, "torchair_graph_config": {"enabled": false, "enable_multistream_shared_expert": false}}'
+    benchmarks:
--- a/tests/e2e/nightly/single_node/models/configs/DeepSeek-V3.2-W8A8.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/DeepSeek-V3.2-W8A8.yaml
@@ -0,0 +1,78 @@
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "DeepSeek-V3.2-W8A8-TP8-DP2"
+    model: "vllm-ascend/DeepSeek-V3.2-W8A8"
+    envs:
+      HCCL_OP_EXPANSION_MODE: "AIV"
+      OMP_PROC_BIND: "false"
+      OMP_NUM_THREADS: "1"
+      HCCL_BUFFSIZE: "1024"
+      VLLM_ASCEND_ENABLE_MLAPO: "1"
+      PYTORCH_NPU_ALLOC_CONF: "expandable_segments:True"
+      VLLM_ASCEND_ENABLE_FLASHCOMM1: "1"
+      VLLM_ENGINE_READY_TIMEOUT_S: "1800"
+      SERVER_PORT: "DEFAULT_PORT"
+    server_cmd:
+      - "--enable-expert-parallel"
+      - "--tensor-parallel-size"
+      - "8"
+      - "--data-parallel-size"
+      - "2"
+      - "--port"
+      - "$SERVER_PORT"
+      - "--max-model-len"
+      - "8192"
+      - "--max-num-batched-tokens"
+      - "8192"
+      - "--max-num-seqs"
+      - "4"
+      - "--trust-remote-code"
+      - "--quantization"
+      - "ascend"
+      - "--gpu-memory-utilization"
+      - "0.98"
+      - "--compilation-config"
+      - '{"cudagraph_capture_sizes":[8, 16, 24, 32, 40, 48], "cudagraph_mode":"FULL_DECODE_ONLY"}'
+      - "--speculative-config"
+      - '{"num_speculative_tokens": 3, "method":"deepseek_mtp"}'
+      - "--additional-config"
+      - '{"layer_sharding": ["q_b_proj", "o_proj"]}'
+      - "--reasoning-parser"
+      - "deepseek_v3"
+      - "--tokenizer_mode"
+      - "deepseek_v32"
+    benchmarks:
+      acc:
+        case_type: accuracy
+        dataset_path: vllm-ascend/gsm8k-lite
+        request_conf: vllm_api_general_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_chat_prompt
+        max_out_len: 4096
+        batch_size: 8
+        baseline: 95
+        threshold: 5
+      perf_1:
+        case_type: performance
+        dataset_path: vllm-ascend/GSM8K-in3500-bs400
+        request_conf: vllm_api_stream_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+        num_prompts: 1
+        max_out_len: 1500
+        batch_size: 1
+        request_rate: 11.2
+        baseline: 134
+        threshold: 0.97
+      perf_2:
+        case_type: performance
+        dataset_path: vllm-ascend/GSM8K-in3500-bs400
+        request_conf: vllm_api_stream_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+        num_prompts: 100
+        max_out_len: 1500
+        batch_size: 4
+        request_rate: 11.2
+        baseline: 134
+        threshold: 0.97
--- a/tests/e2e/nightly/single_node/models/configs/GLM-4.5.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/GLM-4.5.yaml
@@ -0,0 +1,72 @@
+# ==========================================
+# Shared Configurations
+# ==========================================
+
+_envs: &envs
+  HCCL_BUFFSIZE: "1024"
+  SERVER_PORT: "DEFAULT_PORT"
+
+_server_cmd: &server_cmd
+  - "--no-enable-prefix-caching"
+  - "--enable-expert-parallel"
+  - "--tensor-parallel-size"
+  - "8"
+  - "--data-parallel-size"
+  - "2"
+  - "--port"
+  - "$SERVER_PORT"
+  - "--max-model-len"
+  - "8192"
+  - "--max-num-batched-tokens"
+  - "8192"
+  - "--block-size"
+  - "16"
+  - "--trust-remote-code"
+  - "--gpu-memory-utilization"
+  - "0.9"
+
+_benchmarks: &benchmarks
+  acc:
+    case_type: accuracy
+    dataset_path: vllm-ascend/gsm8k-lite
+    request_conf: vllm_api_general_chat
+    dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_chat_prompt
+    max_out_len: 4096
+    batch_size: 8
+    baseline: 95
+    threshold: 5
+  perf:
+    case_type: performance
+    dataset_path: vllm-ascend/GSM8K-in3500-bs400
+    request_conf: vllm_api_stream_chat
+    dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+    num_prompts: 16
+    max_out_len: 1500
+    batch_size: 8
+    request_rate: 0
+    baseline: 1
+    threshold: 0.97
+
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "GLM-4.5-TP8-DP2-fullgraph"
+    model: "ZhipuAI/GLM-4.5"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--compilation-config"
+      - '{"cudagraph_capture": [1,2,4,8,16], "cudagraph_model":"FULL_DECODE_ONLY"}'
+    benchmarks:
+      <<: *benchmarks
+
+  - name: "GLM-4.5-TP8-DP2-eager"
+    model: "ZhipuAI/GLM-4.5"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    benchmarks:
+      <<: *benchmarks
--- a/tests/e2e/nightly/single_node/models/configs/Kimi-K2-Thinking.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/Kimi-K2-Thinking.yaml
@@ -0,0 +1,52 @@
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "Kimi-K2-Thinking-TP16-Case"
+    model: "moonshotai/Kimi-K2-Thinking"
+    envs:
+      HCCL_BUFFSIZE: "1024"
+      TASK_QUEUE_ENABLE: "1"
+      OMP_PROC_BIND: "false"
+      HCCL_OP_EXPANSION_MODE: "AIV"
+      PYTORCH_NPU_ALLOC_CONF: "expandable_segments:True"
+      SERVER_PORT: "DEFAULT_PORT"
+    server_cmd:
+      - "--tensor-parallel-size"
+      - "16"
+      - "--port"
+      - "$SERVER_PORT"
+      - "--max-model-len"
+      - "8192"
+      - "--max-num-batched-tokens"
+      - "8192"
+      - "--max-num-seqs"
+      - "12"
+      - "--gpu-memory-utilization"
+      - "0.9"
+      - "--trust-remote-code"
+      - "--enable-expert-parallel"
+      - "--no-enable-prefix-caching"
+    benchmarks:
+      acc:
+        case_type: accuracy
+        dataset_path: vllm-ascend/gsm8k-lite
+        request_conf: vllm_api_general_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_chat_prompt
+        max_out_len: 4096
+        batch_size: 32
+        baseline: 95
+        threshold: 5
+      perf:
+        case_type: performance
+        dataset_path: vllm-ascend/GSM8K-in3500-bs400
+        request_conf: vllm_api_stream_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+        num_prompts: 512
+        max_out_len: 256
+        batch_size: 64
+        trust_remote_code: true
+        request_rate: 11.2
+        baseline: 1
+        threshold: 0.97
--- a/tests/e2e/nightly/single_node/models/configs/MTPX-DeepSeek-R1-0528-W8A8.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/MTPX-DeepSeek-R1-0528-W8A8.yaml
@@ -0,0 +1,90 @@
+# ==========================================
+# Shared Configurations
+# ==========================================
+
+_envs: &envs
+  OMP_NUM_THREADS: "100"
+  OMP_PROC_BIND: "false"
+  HCCL_BUFFSIZE: "1024"
+  VLLM_RPC_TIMEOUT: "3600000"
+  VLLM_EXECUTE_MODEL_TIMEOUT_SECONDS: "3600000"
+  SERVER_PORT: "DEFAULT_PORT"
+
+_server_cmd: &server_cmd
+  - "--quantization"
+  - "ascend"
+  - "--seed"
+  - "1024"
+  - "--no-enable-prefix-caching"
+  - "--data-parallel-size"
+  - "2"
+  - "--tensor-parallel-size"
+  - "8"
+  - "--enable-expert-parallel"
+  - "--port"
+  - "$SERVER_PORT"
+  - "--max-model-len"
+  - "40960"
+  - "--max-num-seqs"
+  - "14"
+  - "--trust-remote-code"
+
+_benchmarks_gsm8k: &benchmarks_gsm8k
+  acc:
+    case_type: accuracy
+    dataset_path: vllm-ascend/gsm8k-lite
+    request_conf: vllm_api_general_chat
+    dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_chat_prompt
+    max_out_len: 32768
+    batch_size: 32
+    baseline: 95
+    threshold: 5
+
+_benchmarks_aime: &benchmarks_aime
+  acc:
+    case_type: accuracy
+    dataset_path: vllm-ascend/aime2024
+    request_conf: vllm_api_general_chat
+    dataset_conf: aime2024/aime2024_gen_0_shot_chat_prompt
+    max_out_len: 32768
+    batch_size: 32
+    baseline: 86.67
+    threshold: 7
+
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "MTPX-DeepSeek-R1-0528-W8A8-mtp2"
+    model: "vllm-ascend/DeepSeek-R1-0528-W8A8"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--max-num-batched-tokens"
+      - "4096"
+      - "--speculative-config"
+      - '{"num_speculative_tokens": 2, "method": "mtp"}'
+      - "--gpu-memory-utilization"
+      - "0.92"
+    benchmarks:
+      <<: *benchmarks_gsm8k
+
+  - name: "MTPX-DeepSeek-R1-0528-W8A8-mtp3"
+    model: "vllm-ascend/DeepSeek-R1-0528-W8A8"
+    envs:
+      <<: *envs
+      HCCL_OP_EXPANSION_MODE: "AIV"
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--max-num-batched-tokens"
+      - "2048"
+      - "--speculative-config"
+      - '{"num_speculative_tokens": 3, "method": "mtp"}'
+      - "--gpu-memory-utilization"
+      - "0.9"
+      - "--compilation-config"
+      - '{"cudagraph_capture_sizes": [56], "cudagraph_mode": "FULL_DECODE_ONLY"}'
+    benchmarks:
+      <<: *benchmarks_aime
--- a/tests/e2e/nightly/single_node/models/configs/Prefix-Cache-DeepSeek-R1-0528-W8A8.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/Prefix-Cache-DeepSeek-R1-0528-W8A8.yaml
@@ -0,0 +1,77 @@
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "prefix-cache-deepseek-r1-0528-w8a8"
+    model: "vllm-ascend/DeepSeek-R1-0528-W8A8"
+    envs:
+      OMP_NUM_THREADS: "10"
+      OMP_PROC_BIND: "false"
+      HCCL_BUFFSIZE: "1024"
+      PYTORCH_NPU_ALLOC_CONF: "expandable_segments:True"
+      SERVER_PORT: "DEFAULT_PORT"
+    server_cmd:
+      - "--quantization"
+      - "ascend"
+      - "--data-parallel-size"
+      - "2"
+      - "--tensor-parallel-size"
+      - "8"
+      - "--enable-expert-parallel"
+      - "--port"
+      - "$SERVER_PORT"
+      - "--seed"
+      - "1024"
+      - "--max-model-len"
+      - "5200"
+      - "--max-num-batched-tokens"
+      - "4096"
+      - "--max-num-seqs"
+      - "16"
+      - "--trust-remote-code"
+      - "--gpu-memory-utilization"
+      - "0.9"
+      - "--additional-config"
+      - '{"enable_weight_nz_layout": true}'
+      - "--speculative-config"
+      - '{"num_speculative_tokens": 1, "method": "mtp"}'
+    test_content:
+      - "benchmark_comparisons"
+    benchmark_comparisons_args:
+      - metric: "TTFT"
+        baseline: "prefix0"
+        target: "prefix75"
+        ratio: 0.8
+        operator: "<"
+    benchmarks:
+      warm_up:
+        case_type: performance
+        dataset_path: vllm-ascend/GSM8K-in1024-bs210
+        request_conf: vllm_api_stream_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+        num_prompts: 210
+        max_out_len: 2
+        batch_size: 1000
+        baseline: 0
+        threshold: 0.97
+      prefix0:
+        case_type: performance
+        dataset_path: vllm-ascend/prefix0-in3500-bs210
+        request_conf: vllm_api_stream_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+        num_prompts: 210
+        max_out_len: 1500
+        batch_size: 18
+        baseline: 1
+        threshold: 0.97
+      prefix75:
+        case_type: performance
+        dataset_path: vllm-ascend/prefix75-in3500-bs210
+        request_conf: vllm_api_stream_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+        num_prompts: 210
+        max_out_len: 1500
+        batch_size: 18
+        baseline: 1
+        threshold: 0.97
--- a/tests/e2e/nightly/single_node/models/configs/Prefix-Cache-Qwen3-32B-Int8.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/Prefix-Cache-Qwen3-32B-Int8.yaml
@@ -0,0 +1,70 @@
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "prefix-cache-qwen3-32b-w8a8"
+    model: "vllm-ascend/Qwen3-32B-W8A8"
+    envs:
+      TASK_QUEUE_ENABLE: "1"
+      HCCL_OP_EXPANSION_MODE: "AIV"
+      SERVER_PORT: "DEFAULT_PORT"
+    server_cmd:
+      - "--quantization"
+      - "ascend"
+      - "--reasoning-parser"
+      - "qwen3"
+      - "--tensor-parallel-size"
+      - "4"
+      - "--port"
+      - "$SERVER_PORT"
+      - "--max-model-len"
+      - "8192"
+      - "--max-num-batched-tokens"
+      - "8192"
+      - "--max-num-seqs"
+      - "256"
+      - "--trust-remote-code"
+      - "--gpu-memory-utilization"
+      - "0.9"
+      - "--additional-config"
+      - '{"enable_weight_nz_layout": true}'
+    test_content:
+      - "benchmark_comparisons"
+    benchmark_comparisons_args:
+      - metric: "TTFT"
+        baseline: "prefix0"
+        target: "prefix75"
+        ratio: 0.8
+        operator: "<"
+    benchmarks:
+      warm_up:
+        case_type: performance
+        dataset_path: vllm-ascend/GSM8K-in1024-bs210
+        request_conf: vllm_api_stream_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+        num_prompts: 210
+        max_out_len: 2
+        batch_size: 1000
+        baseline: 0
+        threshold: 0.97
+      prefix0:
+        case_type: performance
+        dataset_path: vllm-ascend/prefix0-in3500-bs210
+        request_conf: vllm_api_stream_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+        num_prompts: 210
+        max_out_len: 1500
+        batch_size: 48
+        baseline: 1
+        threshold: 0.97
+      prefix75:
+        case_type: performance
+        dataset_path: vllm-ascend/prefix75-in3500-bs210
+        request_conf: vllm_api_stream_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+        num_prompts: 210
+        max_out_len: 1500
+        batch_size: 48
+        baseline: 1
+        threshold: 0.97
--- a/tests/e2e/nightly/single_node/models/configs/QwQ-32B.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/QwQ-32B.yaml
@@ -0,0 +1,78 @@
+# ==========================================
+# Shared Configurations
+# ==========================================
+
+_envs: &envs
+  TASK_QUEUE_ENABLE: "1"
+  OMP_PROC_BIND: "false"
+  HCCL_OP_EXPANSION_MODE: "AIV"
+  VLLM_ASCEND_ENABLE_FLASHCOMM: "1"
+  VLLM_ASCEND_ENABLE_DEBSE_OPTIMIZE: "1"
+  SERVER_PORT: "DEFAULT_PORT"
+
+_server_cmd: &server_cmd
+  - "--tensor-parallel-size"
+  - "4"
+  - "--port"
+  - "$SERVER_PORT"
+  - "--max-model-len"
+  - "36864"
+  - "--max-num-batched-tokens"
+  - "36864"
+  - "--block-size"
+  - "128"
+  - "--trust-remote-code"
+  - "--gpu-memory-utilization"
+  - "0.9"
+  - "--reasoning-parser"
+  - "deepseek_r1"
+  - "--distributed_executor_backend"
+  - "mp"
+  - "--additional-config"
+  - '{"weight_prefetch_config":{"enabled":true}}'
+
+_benchmarks: &benchmarks
+  acc:
+    case_type: accuracy
+    dataset_path: vllm-ascend/gsm8k-lite
+    request_conf: vllm_api_general_chat
+    dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_chat_prompt
+    max_out_len: 32768
+    batch_size: 32
+    baseline: 95
+    threshold: 5
+  perf:
+    case_type: performance
+    dataset_path: vllm-ascend/GSM8K-in3500-bs400
+    request_conf: vllm_api_stream_chat
+    dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+    num_prompts: 240
+    max_out_len: 1500
+    batch_size: 60
+    baseline: 1
+    threshold: 0.97
+
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "QwQ-32B-aclgraph"
+    model: "Qwen/QwQ-32B"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--compilation_config"
+      - '{"cudagraph_mode":"FULL_DECODE_ONLY", "cudagraph_capture_sizes": [1, 8, 24, 48, 60]}'
+    benchmarks:
+      <<: *benchmarks
+
+  - name: "QwQ-32B-single"
+    model: "Qwen/QwQ-32B"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--enforce-eager"
+    benchmarks:
--- a/tests/e2e/nightly/single_node/models/configs/Qwen2.5-VL-32B-Instruct.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/Qwen2.5-VL-32B-Instruct.yaml
@@ -0,0 +1,63 @@
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "Qwen2.5-VL-32B-Instruct-a3"
+    model: "Qwen/Qwen2.5-VL-32B-Instruct"
+    envs:
+      TASK_QUEUE_ENABLE: "1"
+      VLLM_ASCEND_ENABLE_NZ: "0"
+      HCCL_OP_EXPANSION_MODE: "AIV"
+      SERVER_PORT: "DEFAULT_PORT"
+    server_cmd:
+      - "--no-enable-prefix-caching"
+      - "--mm-processor-cache-gb"
+      - "0"
+      - "--tensor-parallel-size"
+      - "4"
+      - "--port"
+      - "$SERVER_PORT"
+      - "--max-model-len"
+      - "30000"
+      - "--max-num-batched-tokens"
+      - "40000"
+      - "--max-num-seqs"
+      - "400"
+      - "--trust-remote-code"
+      - "--gpu-memory-utilization"
+      - "0.8"
+      - "--compilation_config"
+      - '{"cudagraph_mode": "FULL_DECODE_ONLY"}'
+    test_content:
+      - "completion"
+      - "image"
+    benchmarks:
+      acc:
+        case_type: accuracy
+        dataset_path: vllm-ascend/textvqa-lite
+        request_conf: vllm_api_stream_chat
+        dataset_conf: textvqa/textvqa_gen_base64
+        max_out_len: 2048
+        batch_size: 128
+        baseline: 76.22
+        temperature: 0
+        top_k: -1
+        top_p: 1
+        repetition_penalty: 1
+        threshold: 5
+      perf:
+        case_type: performance
+        dataset_path: vllm-ascend/textvqa-perf-1080p
+        request_conf: vllm_api_stream_chat
+        dataset_conf: textvqa/textvqa_gen_base64
+        num_prompts: 512
+        max_out_len: 256
+        batch_size: 128
+        temperature: 0
+        top_k: -1
+        top_p: 1
+        repetition_penalty: 1
+        request_rate: 0
+        baseline: 1
+        threshold: 0.97
--- a/tests/e2e/nightly/single_node/models/configs/Qwen2.5-VL-7B-Instruct-EPD.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/Qwen2.5-VL-7B-Instruct-EPD.yaml
@@ -0,0 +1,92 @@
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "Qwen2.5-VL-7B-Instruct-epd"
+    model: "Qwen/Qwen2.5-VL-7B-Instruct"
+    service_mode: "epd"
+    envs:
+      ENCODE_PORT: "DEFAULT_PORT"
+      PD_PORT: "DEFAULT_PORT"
+      PROXY_PORT: "DEFAULT_PORT"
+    epd_server_cmds:
+      - - "--port"
+        - "$ENCODE_PORT"
+        - "--model"
+        - "Qwen/Qwen2.5-VL-7B-Instruct"
+        - "--gpu-memory-utilization"
+        - "0.01"
+        - "--tensor-parallel-size"
+        - "1"
+        - "--enforce-eager"
+        - "--no-enable-prefix-caching"
+        - "--max-model-len"
+        - "10000"
+        - "--max-num-batched-tokens"
+        - "10000"
+        - "--max-num-seqs"
+        - "1"
+        - "--ec-transfer-config"
+        - '{"ec_connector_extra_config":{"shared_storage_path":"/dev/shm/epd/storage"},"ec_connector":"ECExampleConnector","ec_role": "ec_producer"}'
+      - - "--port"
+        - "$PD_PORT"
+        - "--model"
+        - "Qwen/Qwen2.5-VL-7B-Instruct"
+        - "--gpu-memory-utilization"
+        - "0.95"
+        - "--tensor-parallel-size"
+        - "1"
+        - "--enforce-eager"
+        - "--max-model-len"
+        - "10000"
+        - "--max-num-batched-tokens"
+        - "10000"
+        - "--max-num-seqs"
+        - "128"
+        - "--ec-transfer-config"
+        - '{"ec_connector_extra_config":{"shared_storage_path":"/dev/shm/epd/storage"},"ec_connector":"ECExampleConnector","ec_role": "ec_consumer"}'
+    epd_proxy_args:
+      - "--host"
+      - "127.0.0.1"
+      - "--port"
+      - "$PROXY_PORT"
+      - "--encode-servers-urls"
+      - "http://localhost:$ENCODE_PORT"
+      - "--decode-servers-urls"
+      - "http://localhost:$PD_PORT"
+      - "--prefill-servers-urls"
+      - "disable"
+    test_content:
+    benchmarks:
+      warm_up:
+        case_type: performance
+        dataset_path: vllm-ascend/textvqa-perf-1080p
+        request_conf: vllm_api_stream_chat
+        dataset_conf: textvqa/textvqa_gen_base64
+        num_prompts: 50
+        max_out_len: 20
+        batch_size: 32
+        request_rate: 0
+        baseline: 1
+        threshold: 0.97
+      acc:
+        case_type: accuracy
+        dataset_path: vllm-ascend/textvqa-lite
+        request_conf: vllm_api_stream_chat
+        dataset_conf: textvqa/textvqa_gen_base64
+        max_out_len: 2048
+        batch_size: 128
+        baseline: 82.05
+        threshold: 5
+      perf:
+        case_type: performance
+        dataset_path: vllm-ascend/textvqa-perf-1080p
+        request_conf: vllm_api_stream_chat
+        dataset_conf: textvqa/textvqa_gen_base64
+        num_prompts: 512
+        max_out_len: 256
+        batch_size: 128
+        request_rate: 0
+        baseline: 1
+        threshold: 0.97
--- a/tests/e2e/nightly/single_node/models/configs/Qwen2.5-VL-7B-Instruct.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/Qwen2.5-VL-7B-Instruct.yaml
@@ -0,0 +1,55 @@
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "Qwen2.5-VL-7B-Instruct"
+    model: "Qwen/Qwen2.5-VL-7B-Instruct"
+    envs:
+      TASK_QUEUE_ENABLE: "1"
+      VLLM_ASCEND_ENABLE_NZ: "0"
+      HCCL_OP_EXPANSION_MODE: "AIV"
+      SERVER_PORT: "DEFAULT_PORT"
+    server_cmd:
+      - "--no-enable-prefix-caching"
+      - "--mm-processor-cache-gb"
+      - "0"
+      - "--tensor-parallel-size"
+      - "4"
+      - "--port"
+      - "$SERVER_PORT"
+      - "--max-model-len"
+      - "30000"
+      - "--max-num-batched-tokens"
+      - "40000"
+      - "--max-num-seqs"
+      - "400"
+      - "--trust-remote-code"
+      - "--gpu-memory-utilization"
+      - "0.8"
+      - "--compilation_config"
+      - '{"cudagraph_mode": "FULL_DECODE_ONLY"}'
+    test_content:
+      - "completion"
+      - "image"
+    benchmarks:
+      acc:
+        case_type: accuracy
+        dataset_path: vllm-ascend/textvqa-lite
+        request_conf: vllm_api_stream_chat
+        dataset_conf: textvqa/textvqa_gen_base64
+        max_out_len: 2048
+        batch_size: 128
+        baseline: 82.05
+        threshold: 5
+      perf:
+        case_type: performance
+        dataset_path: vllm-ascend/textvqa-perf-1080p
+        request_conf: vllm_api_stream_chat
+        dataset_conf: textvqa/textvqa_gen_base64
+        num_prompts: 512
+        max_out_len: 256
+        batch_size: 128
+        request_rate: 0
+        baseline: 1
+        threshold: 0.97
--- a/tests/e2e/nightly/single_node/models/configs/Qwen3-235B-A22B-W8A8.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/Qwen3-235B-A22B-W8A8.yaml
@@ -0,0 +1,85 @@
+# ==========================================
+# Shared Configurations
+# ==========================================
+
+_envs: &envs
+  OMP_NUM_THREADS: "10"
+  OMP_PROC_BIND: "false"
+  HCCL_BUFFSIZE: "1024"
+  PYTORCH_NPU_ALLOC_CONF: "expandable_segments:True"
+  VLLM_ASCEND_ENABLE_FLASHCOMM1: "1"
+  SERVER_PORT: "DEFAULT_PORT"
+
+_server_cmd: &server_cmd
+  - "--quantization"
+  - "ascend"
+  - "--async-scheduling"
+  - "--data-parallel-size"
+  - "4"
+  - "--tensor-parallel-size"
+  - "4"
+  - "--enable-expert-parallel"
+  - "--port"
+  - "$SERVER_PORT"
+  - "--max-model-len"
+  - "40960"
+  - "--max-num-batched-tokens"
+  - "8192"
+  - "--max-num-seqs"
+  - "12"
+  - "--trust-remote-code"
+  - "--gpu-memory-utilization"
+  - "0.9"
+
+_benchmarks: &benchmarks
+  acc:
+    case_type: accuracy
+    dataset_path: vllm-ascend/gsm8k-lite
+    request_conf: vllm_api_general_chat
+    dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_chat_prompt
+    max_out_len: 32768
+    batch_size: 32
+    top_k: 20
+    baseline: 95
+    threshold: 5
+
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "Qwen3-235B-A22B-W8A8-full_graph"
+    model: "vllm-ascend/Qwen3-235B-A22B-W8A8"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--compilation-config"
+      - '{"cudagraph_mode": "FULL_DECODE_ONLY"}'
+    benchmarks:
+      <<: *benchmarks
+
+  - name: "Qwen3-235B-A22B-W8A8-piecewise"
+    model: "vllm-ascend/Qwen3-235B-A22B-W8A8"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--compilation-config"
+      - '{"cudagraph_mode": "PIECEWISE"}'
+    benchmarks:
+      <<: *benchmarks
+
+  - name: "Qwen3-235B-A22B-W8A8-EPLB"
+    model: "vllm-ascend/Qwen3-235B-A22B-W8A8"
+    envs:
+      <<: *envs
+      DYNAMIC_EPLB: "true"
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--additional-config"
+      - '{"eplb_config": {"dynamic_eplb": "true", "expert_heat_collection_interval": 600, "algorithm_execution_interval": 50, "num_redundant_experts": 16, "eplb_policy_type": 2}}'
+      - "--compilation-config"
+      - '{"cudagraph_mode": "FULL_DECODE_ONLY"}'
+    benchmarks:
+      <<: *benchmarks
--- a/tests/e2e/nightly/single_node/models/configs/Qwen3-30B-A3B-W8A8.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/Qwen3-30B-A3B-W8A8.yaml
@@ -0,0 +1,46 @@
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "Qwen3-30B-A3B-W8A8-TP1"
+    model: "vllm-ascend/Qwen3-30B-A3B-W8A8"
+    envs:
+      OMP_PROC_BIND: "false"
+      OMP_NUM_THREADS: "10"
+      HCCL_BUFFSIZE: "1024"
+      HCCL_OP_EXPANSION_MODE: "AIV"
+      PYTORCH_NPU_ALLOC_CONF: "expandable_segments:True"
+      SERVER_PORT: "DEFAULT_PORT"
+    server_cmd:
+      - "--quantization"
+      - "ascend"
+      - "--async-scheduling"
+      - "--no-enable-prefix-caching"
+      - "--tensor-parallel-size"
+      - "1"
+      - "--port"
+      - "$SERVER_PORT"
+      - "--max-model-len"
+      - "5600"
+      - "--max-num-batched-tokens"
+      - "16384"
+      - "--max-num-seqs"
+      - "100"
+      - "--trust-remote-code"
+      - "--gpu-memory-utilization"
+      - "0.9"
+      - "--compilation-config"
+      - '{"cudagraph_mode": "FULL_DECODE_ONLY"}'
+    benchmarks:
+      perf:
+        case_type: performance
+        dataset_path: vllm-ascend/GSM8K-in3500-bs400
+        request_conf: vllm_api_stream_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+        num_prompts: 180
+        max_out_len: 1500
+        batch_size: 45
+        request_rate: 0
+        baseline: 1
+        threshold: 0.97
--- a/tests/e2e/nightly/single_node/models/configs/Qwen3-32B-Int8-A2.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/Qwen3-32B-Int8-A2.yaml
@@ -0,0 +1,79 @@
+# ==========================================
+# Shared Configurations
+# ==========================================
+
+_envs: &envs
+  TASK_QUEUE_ENABLE: "1"
+  HCCL_OP_EXPANSION_MODE: "AIV"
+  VLLM_ASCEND_ENABLE_FLASHCOMM: "1"
+  SERVER_PORT: "DEFAULT_PORT"
+
+_server_cmd: &server_cmd
+  - "--quantization"
+  - "ascend"
+  - "--no-enable-prefix-caching"
+  - "--tensor-parallel-size"
+  - "4"
+  - "--port"
+  - "$SERVER_PORT"
+  - "--max-model-len"
+  - "40960"
+  - "--max-num-batched-tokens"
+  - "40960"
+  - "--block-size"
+  - "128"
+  - "--trust-remote-code"
+  - "--reasoning-parser"
+  - "qwen3"
+  - "--gpu-memory-utilization"
+  - "0.9"
+  - "--async-scheduling"
+  - "--additional-config"
+  - '{"weight_prefetch_config":{"enabled":true}}'
+
+_benchmarks: &benchmarks
+  acc:
+    case_type: accuracy
+    dataset_path: vllm-ascend/aime2024
+    request_conf: vllm_api_general_chat
+    dataset_conf: aime2024/aime2024_gen_0_shot_chat_prompt
+    max_out_len: 32768
+    batch_size: 32
+    baseline: 83.33
+    threshold: 7
+
+  perf:
+    case_type: performance
+    dataset_path: vllm-ascend/GSM8K-in3500-bs400
+    request_conf: vllm_api_stream_chat
+    dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+    num_prompts: 288
+    max_out_len: 1500
+    batch_size: 72
+    baseline: 1
+    threshold: 0.97
+
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "Qwen3-32B-W8A8-aclgraph-a2"
+    model: "vllm-ascend/Qwen3-32B-W8A8"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--compilation-config"
+      - '{"cudagraph_mode":"FULL_DECODE_ONLY","cudagraph_capture_sizes":[1,12,16,20,24,32,48,60,64,68,72,76,80]}'
+    benchmarks:
+      <<: *benchmarks
+
+  - name: "Qwen3-32B-W8A8-single-a2"
+    model: "vllm-ascend/Qwen3-32B-W8A8"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--enforce-eager"
+    benchmarks:
--- a/tests/e2e/nightly/single_node/models/configs/Qwen3-32B-Int8-A3-Feature-Stack3.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/Qwen3-32B-Int8-A3-Feature-Stack3.yaml
@@ -0,0 +1,69 @@
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "Qwen3-32B-W8A8-a3-feature-stack3"
+    model: "vllm-ascend/Qwen3-32B-W8A8"
+    envs:
+      VLLM_USE: "1"
+      TASK_QUEUE_ENABLE: "1"
+      HCCL_OP_EXPANSION_MODE: "AIV"
+      OMP_PROC_BIND: "false"
+      VLLM_ASCEND_ENABLE_TOPK_OPTIMIZE: "1"
+      VLLM_ASCEND_ENABLE_FLASHCOMM: "1"
+      SERVER_PORT: "DEFAULT_PORT"
+    prompts:
+      - "9.11 and 9.8, which is greater?"
+    api_keyword_args:
+      chat_template_kwargs:
+        enable_thinking: true
+    server_cmd:
+      - "--quantization"
+      - "ascend"
+      - "--tensor-parallel-size"
+      - "4"
+      - "--port"
+      - "$SERVER_PORT"
+      - "--trust-remote-code"
+      - "--reasoning-parser"
+      - "qwen3"
+      - "--distributed_executor_backend"
+      - "mp"
+      - "--gpu-memory-utilization"
+      - "0.9"
+      - "--block-size"
+      - "128"
+      - "--max-num-seqs"
+      - "256"
+      - "--enforce-eager"
+      - "--max-model-len"
+      - "35840"
+      - "--max-num-batched-tokens"
+      - "35840"
+      - "--additional-config"
+      - '{"enable_weight_nz_layout":true, "weight_prefetch_config":{"enabled": true}}'
+      - "--compilation-config"
+      - '{"cudagraph_mode":"FULL_DECODE_ONLY", "cudagraph_capture_sizes":[1,8,24,48,60]}'
+    test_content:
+      - "chat_completion"
+    benchmarks:
+      acc:
+        case_type: accuracy
+        dataset_path: vllm-ascend/gsm8k-lite
+        request_conf: vllm_api_general_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_noncot_chat_prompt
+        max_out_len: 10240
+        batch_size: 32
+        baseline: 96
+        threshold: 4
+      perf:
+        case_type: performance
+        dataset_path: vllm-ascend/GSM8K-in3500-bs400
+        request_conf: vllm_api_stream_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+        num_prompts: 240
+        max_out_len: 1500
+        batch_size: 60
+        baseline: 1
+        threshold: 0.97
--- a/tests/e2e/nightly/single_node/models/configs/Qwen3-32B-Int8.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/Qwen3-32B-Int8.yaml
@@ -0,0 +1,78 @@
+# ==========================================
+# Shared Configurations
+# ==========================================
+
+_envs: &envs
+  TASK_QUEUE_ENABLE: "1"
+  HCCL_OP_EXPANSION_MODE: "AIV"
+  VLLM_ASCEND_ENABLE_FLASHCOMM: "1"
+  SERVER_PORT: "DEFAULT_PORT"
+
+_server_cmd: &server_cmd
+  - "--quantization"
+  - "ascend"
+  - "--no-enable-prefix-caching"
+  - "--tensor-parallel-size"
+  - "4"
+  - "--port"
+  - "$SERVER_PORT"
+  - "--max-model-len"
+  - "40960"
+  - "--max-num-batched-tokens"
+  - "40960"
+  - "--block-size"
+  - "128"
+  - "--trust-remote-code"
+  - "--reasoning-parser"
+  - "qwen3"
+  - "--gpu-memory-utilization"
+  - "0.9"
+  - "--async-scheduling"
+  - "--additional-config"
+  - '{"weight_prefetch_config":{"enabled":true}}'
+
+_benchmarks: &benchmarks
+  acc:
+    case_type: accuracy
+    dataset_path: vllm-ascend/aime2024
+    request_conf: vllm_api_general_chat
+    dataset_conf: aime2024/aime2024_gen_0_shot_chat_prompt
+    max_out_len: 32768
+    batch_size: 32
+    baseline: 83.33
+    threshold: 7
+  perf:
+        case_type: performance
+        dataset_path: vllm-ascend/GSM8K-in3500-bs400
+        request_conf: vllm_api_stream_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+        num_prompts: 304
+        max_out_len: 1500
+        batch_size: 76
+        baseline: 1
+        threshold: 0.97
+
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "Qwen3-32B-W8A8-aclgraph-a3"
+    model: "vllm-ascend/Qwen3-32B-W8A8"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--compilation-config"
+      - '{"cudagraph_mode":"FULL_DECODE_ONLY","cudagraph_capture_sizes":[1,12,16,20,24,32,48,60,64,68,72,76,80]}'
+    benchmarks:
+      <<: *benchmarks
+
+  - name: "Qwen3-32B-W8A8-single-a3"
+    model: "vllm-ascend/Qwen3-32B-W8A8"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--enforce-eager"
+    benchmarks:
--- a/tests/e2e/nightly/single_node/models/configs/Qwen3-32B.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/Qwen3-32B.yaml
@@ -0,0 +1,51 @@
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "Qwen3-32B-TP4"
+    model: "Qwen/Qwen3-32B"
+    envs:
+      TASK_QUEUE_ENABLE: "1"
+      OMP_PROC_BIND: "false"
+      HCCL_OP_EXPANSION_MODE: "AIV"
+      PAGED_ATTENTION_MASK_LEN: "5500"
+      SERVER_PORT: "DEFAULT_PORT"
+    server_cmd:
+      - "--no-enable-prefix-caching"
+      - "--tensor-parallel-size"
+      - "4"
+      - "--port"
+      - "$SERVER_PORT"
+      - "--max-model-len"
+      - "36864"
+      - "--max-num-batched-tokens"
+      - "36864"
+      - "--block-size"
+      - "128"
+      - "--trust-remote-code"
+      - "--gpu-memory-utilization"
+      - "0.9"
+      - "--additional-config"
+      - '{"enable_weight_nz_layout":true}'
+    benchmarks:
+      acc:
+        case_type: accuracy
+        dataset_path: vllm-ascend/gsm8k-lite
+        request_conf: vllm_api_general_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_chat_prompt
+        max_out_len: 32768
+        batch_size: 32
+        baseline: 95
+        threshold: 5
+      perf:
+        case_type: performance
+        dataset_path: vllm-ascend/GSM8K-in3500-bs400
+        request_conf: vllm_api_stream_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+        num_prompts: 80
+        max_out_len: 1500
+        batch_size: 20
+        request_rate: 0
+        baseline: 1
+        threshold: 0.97
--- a/tests/e2e/nightly/single_node/models/configs/Qwen3-Next-80B-A3B-Instruct-A2.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/Qwen3-Next-80B-A3B-Instruct-A2.yaml
@@ -0,0 +1,75 @@
+# ==========================================
+# Shared Configurations
+# ==========================================
+
+_envs: &envs
+  OMP_NUM_THREADS: "10"
+  OMP_PROC_BIND: "false"
+  HCCL_BUFFSIZE: "1024"
+  PYTORCH_NPU_ALLOC_CONF: "expandable_segments:True"
+  SERVER_PORT: "DEFAULT_PORT"
+
+_server_cmd: &server_cmd
+  - "--tensor-parallel-size"
+  - "4"
+  - "--port"
+  - "$SERVER_PORT"
+  - "--max-model-len"
+  - "40960"
+  - "--trust-remote-code"
+  - "--async-scheduling"
+  - "--no-enable-prefix-caching"
+  - "--enable-expert-parallel"
+  - "--gpu-memory-utilization"
+  - "0.8"
+  - "--max-num-seqs"
+  - "64"
+
+_benchmarks: &benchmarks
+  perf:
+    case_type: performance
+    dataset_path: vllm-ascend/GSM8K-in3500-bs400
+    request_conf: vllm_api_stream_chat
+    dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+    num_prompts: 256
+    max_out_len: 1500
+    batch_size: 64
+    baseline: 1
+    threshold: 0.97
+  acc:
+    case_type: accuracy
+    dataset_path: vllm-ascend/gsm8k-lite
+    request_conf: vllm_api_general_chat
+    dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_chat_prompt
+    max_out_len: 32768
+    batch_size: 32
+    top_k: 20
+    baseline: 95
+    threshold: 5
+
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "Qwen3-Next-80B-A3B-Instruct-aclgraph-8192-a2"
+    model: "Qwen/Qwen3-Next-80B-A3B-Instruct"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--max-num-batched-tokens"
+      - "8192"
+    benchmarks:
+      <<: *benchmarks
+
+  - name: "Qwen3-Next-80B-A3B-Instruct-aclgraph-32768-a2"
+    model: "Qwen/Qwen3-Next-80B-A3B-Instruct"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--max-num-batched-tokens"
+      - "32768"
+    benchmarks:
+      <<: *benchmarks
--- a/tests/e2e/nightly/single_node/models/configs/Qwen3-Next-80B-A3B-Instruct-W8A8.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/Qwen3-Next-80B-A3B-Instruct-W8A8.yaml
@@ -0,0 +1,45 @@
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "Qwen3-Next-80B-A3B-Instruct-W8A8"
+    model: "vllm-ascend/Qwen3-Next-80B-A3B-Instruct-W8A8"
+    envs:
+      OMP_NUM_THREADS: "10"
+      OMP_PROC_BIND: "false"
+      HCCL_BUFFSIZE: "1024"
+      SERVER_PORT: "DEFAULT_PORT"
+    server_cmd:
+      - "--quantization"
+      - "ascend"
+      - "--async-scheduling"
+      - "--no-enable-prefix-caching"
+      - "--data-parallel-size"
+      - "1"
+      - "--tensor-parallel-size"
+      - "4"
+      - "--enable-expert-parallel"
+      - "--port"
+      - "$SERVER_PORT"
+      - "--max-model-len"
+      - "40960"
+      - "--max-num-batched-tokens"
+      - "8192"
+      - "--max-num-seqs"
+      - "32"
+      - "--trust-remote-code"
+      - "--gpu-memory-utilization"
+      - "0.65"
+      - "--compilation-config"
+      - '{"cudagraph_capture_sizes": [32]}'
+    benchmarks:
+      acc:
+        case_type: accuracy
+        dataset_path: vllm-ascend/gsm8k-lite
+        request_conf: vllm_api_general_chat
+        dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_chat_prompt
+        max_out_len: 32768
+        batch_size: 32
+        baseline: 95
+        threshold: 5
--- a/tests/e2e/nightly/single_node/models/configs/Qwen3-Next-80B-A3B-Instruct.yaml
+++ b/tests/e2e/nightly/single_node/models/configs/Qwen3-Next-80B-A3B-Instruct.yaml
@@ -0,0 +1,75 @@
+# ==========================================
+# Shared Configurations
+# ==========================================
+
+_envs: &envs
+  OMP_NUM_THREADS: "10"
+  OMP_PROC_BIND: "false"
+  HCCL_BUFFSIZE: "1024"
+  PYTORCH_NPU_ALLOC_CONF: "expandable_segments:True"
+  SERVER_PORT: "DEFAULT_PORT"
+
+_server_cmd: &server_cmd
+  - "--tensor-parallel-size"
+  - "4"
+  - "--port"
+  - "$SERVER_PORT"
+  - "--max-model-len"
+  - "40960"
+  - "--trust-remote-code"
+  - "--async-scheduling"
+  - "--no-enable-prefix-caching"
+  - "--enable-expert-parallel"
+  - "--gpu-memory-utilization"
+  - "0.8"
+  - "--max-num-seqs"
+  - "64"
+
+_benchmarks: &benchmarks
+  perf:
+    case_type: performance
+    dataset_path: vllm-ascend/GSM8K-in3500-bs400
+    request_conf: vllm_api_stream_chat
+    dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_str_perf
+    num_prompts: 256
+    max_out_len: 1500
+    batch_size: 64
+    baseline: 1
+    threshold: 0.97
+  acc:
+    case_type: accuracy
+    dataset_path: vllm-ascend/gsm8k-lite
+    request_conf: vllm_api_general_chat
+    dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_chat_prompt
+    max_out_len: 32768
+    batch_size: 32
+    top_k: 20
+    baseline: 95
+    threshold: 5
+
+# ==========================================
+# ACTUAL TEST CASES
+# ==========================================
+
+test_cases:
+  - name: "Qwen3-Next-80B-A3B-Instruct-aclgraph-8192-a3"
+    model: "Qwen/Qwen3-Next-80B-A3B-Instruct"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--max-num-batched-tokens"
+      - "8192"
+    benchmarks:
+      <<: *benchmarks
+
+  - name: "Qwen3-Next-80B-A3B-Instruct-aclgraph-32768-a3"
+    model: "Qwen/Qwen3-Next-80B-A3B-Instruct"
+    envs:
+      <<: *envs
+    server_cmd: *server_cmd
+    server_cmd_extra:
+      - "--max-num-batched-tokens"
+      - "32768"
+    benchmarks:
+      <<: *benchmarks