diff --git a/.github/workflows/llm_performance_tests.yml b/.github/workflows/llm_performance_tests.yml
index 2d0a293e459..cc96a9c8c92 100644
--- a/.github/workflows/llm_performance_tests.yml
+++ b/.github/workflows/llm_performance_tests.yml
@@ -120,6 +120,7 @@ jobs:
           cd python/llm/dev/benchmark/all-in-one
           mkdir test_batch1
           mkdir test_batch2
+          mkdir test_batch4
           # batch_size 1
           # hide time info
           sed -i 's/str(end - st)/"xxxxxx"/g' run.py
@@ -135,6 +136,14 @@ jobs:
           sed -i 's/batch1/batch2/g' run.py
           python run.py
           mv *.csv test_batch2
+          # batch_size 4
+          cd ../../../../../ 
+          cp python/llm/test/benchmark/arc-perf-test-batch4.yaml python/llm/dev/benchmark/all-in-one/config.yaml
+          cd python/llm/dev/benchmark/all-in-one
+          # change csv name
+          sed -i 's/batch2/batch4/g' run.py
+          python run.py
+          mv *.csv test_batch4
 
       - name: Test on xpu(transformers==4.37.0)
         shell: bash
@@ -148,7 +157,7 @@ jobs:
           cp python/llm/test/benchmark/arc-perf-transformers-437.yaml python/llm/dev/benchmark/all-in-one/config.yaml
           cd python/llm/dev/benchmark/all-in-one
           # change csv name
-          sed -i 's/test1_batch2/test2_batch1/g' run.py
+          sed -i 's/test1_batch4/test2_batch1/g' run.py
           python run.py
           mv *.csv test_batch1
           # batch_size 2
@@ -159,6 +168,14 @@ jobs:
           sed -i 's/batch1/batch2/g' run.py
           python run.py
           mv *.csv test_batch2
+          # batch_size 4
+          cd ../../../../../ 
+          cp python/llm/test/benchmark/arc-perf-transformers-437-batch4.yaml python/llm/dev/benchmark/all-in-one/config.yaml
+          cd python/llm/dev/benchmark/all-in-one
+          # change csv name 
+          sed -i 's/batch2/batch4/g' run.py
+          python run.py
+          mv *.csv test_batch4
 
       - name: Concat csv and generate html
         shell: bash
@@ -185,6 +202,17 @@ jobs:
           done
           cd ../../../../test/benchmark
           python csv_to_html.py -f $CSV_SAVE_PATH/batch_size_2
+          # batch_size 4
+          cd ../../../../
+          cd python/llm/dev/benchmark/all-in-one/test_batch4
+          python ../../../../test/benchmark/concat_csv.py
+          for file in *.csv; do
+              if [[ $file != *test* ]]; then
+                  cp "$file" $CSV_SAVE_PATH/batch_size_4
+              fi
+          done
+          cd ../../../../test/benchmark
+          python csv_to_html.py -f $CSV_SAVE_PATH/batch_size_4
       
       - name: Merge and sort csv files of multiple batches and generate html 
         shell: bash
@@ -204,6 +232,12 @@ jobs:
                   cp "$file" ../../../../test/benchmark/merged_temp
               fi
           done
+          cd ../test_batch4
+          for file in *.csv; do
+              if [[ $file != *test* ]]; then
+                  cp "$file" ../../../../test/benchmark/merged_temp
+              fi
+          done
           cd ../../../../test/benchmark
           python merge_csv_batch.py -f ./merged_temp
           cd merged_temp
@@ -244,6 +278,16 @@ jobs:
           fi
           cd ../
           rm -r test_batch2
+          # batch_size 4
+          cd test_batch4
+          python ../../../../test/benchmark/check_results.py -c test1 -y ../../../../test/benchmark/arc-perf-test-batch4.yaml
+          python ../../../../test/benchmark/check_results.py -c test2 -y ../../../../test/benchmark/arc-perf-transformers-437-batch4.yaml
+          find . -name "*test*.csv" -delete
+          if [ ${{ github.event_name }} == "schedule" ] || [ ${{ github.event_name }} == "workflow_dispatch" ]; then
+            curl -T ./*.csv ${LLM_FTP_URL}/llm/nightly_perf/gpu/
+          fi
+          cd ../
+          rm -r test_batch4
           
           
   llm-performance-test-on-spr:
diff --git a/python/llm/test/benchmark/arc-perf-test-batch4.yaml b/python/llm/test/benchmark/arc-perf-test-batch4.yaml
new file mode 100644
index 00000000000..b74d3833387
--- /dev/null
+++ b/python/llm/test/benchmark/arc-perf-test-batch4.yaml
@@ -0,0 +1,55 @@
+repo_id:
+  - 'meta-llama/Llama-2-7b-chat-hf'
+  - 'meta-llama/Llama-2-13b-chat-hf'
+  - 'THUDM/chatglm2-6b'
+  - 'THUDM/chatglm3-6b-4bit'
+  - 'tiiuae/falcon-7b-instruct-with-patch'
+  - 'mosaicml/mpt-7b-chat'
+  - 'redpajama/gptneox-7b-redpajama-bf16'
+  - 'bigcode/starcoder-15.5b-4bit'
+  - 'databricks/dolly-v1-6b'
+  - 'databricks/dolly-v2-7b'
+  - 'databricks/dolly-v2-12b'
+  - 'internlm/internlm-chat-7b' 
+  - 'Qwen/Qwen-7B-Chat'
+  - 'BAAI/AquilaChat-7B'
+  - 'baichuan-inc/Baichuan2-7B-Chat'
+  - 'baichuan-inc/Baichuan2-13B-Chat-4bit'
+  - 'bigscience/bloomz-7b1'
+#  - 'fnlp/moss-moon-003-sft-4bit' # moss-moon-003-sft cannot work on transformers 4.34+
+  - 'mistralai/Mistral-7B-v0.1' #mwj: need to check
+local_model_hub: '/mnt/disk1/models'
+warm_up: 1
+num_trials: 3
+num_beams: 1 # default to greedy search
+low_bit: 'sym_int4' # default to use 'sym_int4' (i.e. symmetric int4)
+batch_size: 4 # default to 1
+in_out_pairs:
+  - '32-32'
+  - '1024-128'
+  - '2048-256'
+test_api:
+  - "transformer_int4_fp16_gpu"  # on Intel GPU
+cpu_embedding: False # whether put embedding to CPU (only avaiable now for gpu win related test_api)
+exclude:
+  - 'meta-llama/Llama-2-13b-chat-hf:2048'
+  - 'tiiuae/falcon-7b-instruct-with-patch:2048'
+  - 'mosaicml/mpt-7b-chat:2048'
+  - 'redpajama/gptneox-7b-redpajama-bf16:2048'
+  - 'bigcode/starcoder-15.5b-4bit:1024'
+  - 'bigcode/starcoder-15.5b-4bit:2048' 
+  - 'databricks/dolly-v1-6b:2048' 
+  - 'databricks/dolly-v2-7b:2048' 
+  - 'databricks/dolly-v2-12b:1024' 
+  - 'databricks/dolly-v2-12b:2048' 
+  - 'internlm/internlm-chat-7b:2048' 
+  - 'Qwen/Qwen-7B-Chat:2048'
+  - 'BAAI/AquilaChat-7B:2048'
+  - 'baichuan-inc/Baichuan2-7B-Chat:2048'
+  - 'baichuan-inc/Baichuan2-13B-Chat-4bit:1024'
+  - 'baichuan-inc/Baichuan2-13B-Chat-4bit:2048'
+  - 'bigscience/bloomz-7b1:1024'
+  - 'bigscience/bloomz-7b1:2048' 
+#  - 'fnlp/moss-moon-003-sft-4bit:1024'
+#  - 'fnlp/moss-moon-003-sft-4bit:2048'
+task: 'continuation' # task can be 'continuation', 'QA' and 'summarize'
diff --git a/python/llm/test/benchmark/arc-perf-transformers-437-batch4.yaml b/python/llm/test/benchmark/arc-perf-transformers-437-batch4.yaml
new file mode 100644
index 00000000000..e09869164f0
--- /dev/null
+++ b/python/llm/test/benchmark/arc-perf-transformers-437-batch4.yaml
@@ -0,0 +1,23 @@
+# For the models that require transformers 4.37.0
+repo_id:
+  - 'Qwen/Qwen1.5-7B-Chat'
+  - 'microsoft/phi-2'
+  - 'microsoft/Phi-3-mini-4k-instruct'
+  - 'meta-llama/Meta-Llama-3-8B-Instruct' # mwj: need to test
+local_model_hub: '/mnt/disk1/models'
+warm_up: 1
+num_trials: 3
+num_beams: 1 # default to greedy search
+low_bit: 'sym_int4' # default to use 'sym_int4' (i.e. symmetric int4)
+batch_size: 4 # default to 1
+in_out_pairs:
+  - '32-32'
+  - '1024-128'
+  - '2048-256'
+test_api:
+  - "transformer_int4_fp16_gpu"  # on Intel GPU
+cpu_embedding: False # whether put embedding to CPU (only avaiable now for gpu win related test_api)
+exclude:
+  - 'Qwen/Qwen1.5-7B-Chat:2048'
+  - 'meta-llama/Meta-Llama-3-8B-Instruct:2048'
+task: 'continuation' # task can be 'continuation', 'QA' and 'summarize'
\ No newline at end of file
diff --git a/python/llm/test/benchmark/update_html_in_parent_folder.py b/python/llm/test/benchmark/update_html_in_parent_folder.py
index dea237fa926..eb9cc97d493 100644
--- a/python/llm/test/benchmark/update_html_in_parent_folder.py
+++ b/python/llm/test/benchmark/update_html_in_parent_folder.py
@@ -24,13 +24,13 @@
 def update_html_in_parent_folder(folder_path):
     
     current_folder = Path(folder_path)
-    folder_list = [current_folder/'batch_size_1/',current_folder/'batch_size_2/',current_folder/'merged/']
+    folder_list = [current_folder/'batch_size_1/',current_folder/'batch_size_2/',current_folder/'batch_size_4/',current_folder/'merged/']
     
     # List all html files under current folder and delete them
     for html_file in current_folder.glob('*.html'):
         html_file.unlink()
     for folder in folder_list:
-        # Find latest html file under batch1/batch2/merged folders
+        # Find latest html file under batch1/batch2/batch4/merged folders
         latest_html_file = max(Path(folder).glob('*.html'), key=os.path.getctime, default=None)
         # Copy the latest html file to parent folder
         if latest_html_file is not None: