diff --git a/.github/workflows/ci-coverage-overview.yml b/.github/workflows/ci-coverage-overview.yml
index 3fc5512f6..db9269d67 100644
--- a/.github/workflows/ci-coverage-overview.yml
+++ b/.github/workflows/ci-coverage-overview.yml
@@ -6,7 +6,7 @@ on:
   pull_request:
     paths:
       - '.github/workflows/ci-coverage-overview.yml'
-      - 'scripts/ci/ci_coverage_report.py'
+      - 'scripts/ci/utils/ci_coverage_report.py'
       - 'test/registered/**'
   workflow_dispatch:
     inputs:
@@ -34,7 +34,7 @@ jobs:
 
       - name: Generate Summary Report
         run: |
-          python scripts/ci/ci_coverage_report.py --section summary
+          python scripts/ci/utils/ci_coverage_report.py --section summary
 
   by-folder:
     name: Tests by Folder
@@ -50,7 +50,7 @@ jobs:
 
       - name: Generate Tests by Folder Report
         run: |
-          python scripts/ci/ci_coverage_report.py --section by-folder
+          python scripts/ci/utils/ci_coverage_report.py --section by-folder
 
   by-suite:
     name: Tests by Suite
@@ -66,7 +66,7 @@ jobs:
 
       - name: Generate Tests by Suite Report
         run: |
-          python scripts/ci/ci_coverage_report.py --section by-suite
+          python scripts/ci/utils/ci_coverage_report.py --section by-suite
 
   json-export:
     name: JSON Export
@@ -83,7 +83,7 @@ jobs:
 
       - name: Generate JSON Report
         run: |
-          python scripts/ci/ci_coverage_report.py --output-format json > ci_coverage.json
+          python scripts/ci/utils/ci_coverage_report.py --output-format json > ci_coverage.json
 
       - name: Upload JSON artifact
         uses: actions/upload-artifact@v4
diff --git a/.github/workflows/execute-notebook.yml b/.github/workflows/execute-notebook.yml
index 52942c77c..953f34b72 100644
--- a/.github/workflows/execute-notebook.yml
+++ b/.github/workflows/execute-notebook.yml
@@ -24,7 +24,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_dependency.sh
+          bash scripts/ci/cuda/ci_install_dependency.sh
           pip install -r docs/requirements.txt
           apt-get update && apt-get install -y pandoc parallel retry
           ln -sf "$(which python3)" /usr/bin/python
diff --git a/.github/workflows/nightly-test-amd.yml b/.github/workflows/nightly-test-amd.yml
index 936fbb778..1e0a9bf11 100644
--- a/.github/workflows/nightly-test-amd.yml
+++ b/.github/workflows/nightly-test-amd.yml
@@ -71,17 +71,17 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Nightly Unit Test (1-GPU)
         timeout-minutes: 60
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd-1-gpu --nightly --timeout-per-file 600 --continue-on-error || TEST_EXIT_CODE=$?
           echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
@@ -101,17 +101,17 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Nightly Test (2-GPU)
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd --nightly --timeout-per-file 7200 || TEST_EXIT_CODE=$?
           echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
@@ -130,18 +130,18 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Nightly Accuracy Test (2-GPU VLM MMMU)
         timeout-minutes: 180
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd-accuracy-2-gpu-vlm --nightly --timeout-per-file 7200 || TEST_EXIT_CODE=$?
           echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
@@ -160,18 +160,18 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Performance Test (2-GPU Text Models)
         timeout-minutes: 120
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e SGLANG_USE_AITER=1 \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd-perf-text-2-gpu --nightly --timeout-per-file 3600 || TEST_EXIT_CODE=$?
@@ -191,18 +191,18 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Performance Test (2-GPU VLM Models)
         timeout-minutes: 180
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e SGLANG_USE_AITER=1 \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd-perf-vlm-2-gpu --nightly --timeout-per-file 7200 || TEST_EXIT_CODE=$?
@@ -222,17 +222,17 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Accuracy Test (8-GPU GPT-OSS)
         timeout-minutes: 180
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd-accuracy-8-gpu-gpt-oss --nightly --timeout-per-file 7200 || TEST_EXIT_CODE=$?
           echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
@@ -241,7 +241,7 @@ jobs:
       - name: Accuracy Test (8-GPU Grok1-FP8)
         timeout-minutes: 60
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e RCCL_MSCCL_ENABLE=0 \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd-accuracy-8-gpu-grok1-fp8 --nightly --timeout-per-file 3600 || TEST_EXIT_CODE=$?
@@ -261,17 +261,17 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Accuracy Test (8-GPU DeepSeek-R1)
         timeout-minutes: 240
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd-accuracy-8-gpu-deepseek-r1 --nightly --timeout-per-file 7200 || TEST_EXIT_CODE=$?
           echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
@@ -291,18 +291,18 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Accuracy Test (8-GPU Grok1-INT4)
         timeout-minutes: 60
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e RCCL_MSCCL_ENABLE=0 \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd-accuracy-8-gpu-grok1-int4 --nightly --timeout-per-file 3600 || TEST_EXIT_CODE=$?
@@ -314,7 +314,7 @@ jobs:
         continue-on-error: true  # Perf test failure doesn't fail the job if accuracy passed
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e RCCL_MSCCL_ENABLE=0 \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-perf-8-gpu-grok1-int4 --nightly --timeout-per-file 3600 || TEST_EXIT_CODE=$?
@@ -334,18 +334,18 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Accuracy Test (8-GPU Grok2)
         timeout-minutes: 60
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e RCCL_MSCCL_ENABLE=0 \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd-accuracy-8-gpu-grok2 --nightly --timeout-per-file 3600 || TEST_EXIT_CODE=$?
@@ -357,7 +357,7 @@ jobs:
         continue-on-error: true  # Perf test failure doesn't fail the job if accuracy passed
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e RCCL_MSCCL_ENABLE=0 \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-perf-8-gpu-grok2 --nightly --timeout-per-file 3600 || TEST_EXIT_CODE=$?
@@ -377,18 +377,18 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Accuracy Test (8-GPU DeepSeek-V3.1)
         timeout-minutes: 120
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e SGLANG_USE_AITER=1 \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd-accuracy-8-gpu-deepseek-v31 --nightly --timeout-per-file 3600 || TEST_EXIT_CODE=$?
@@ -400,7 +400,7 @@ jobs:
         continue-on-error: true  # Perf test failure doesn't fail the job if accuracy passed
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e SGLANG_USE_ROCM700A=1 \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-perf-8-gpu-deepseek-v31 --nightly --timeout-per-file 18000 || TEST_EXIT_CODE=$?
@@ -421,20 +421,20 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/amd_ci_install_dependency.sh
+          bash scripts/ci/amd/amd_ci_install_dependency.sh
           # Install tabulate for run_suite.py (missing in MI35x container)
-          bash scripts/ci/amd_ci_exec.sh pip install tabulate
+          bash scripts/ci/amd/amd_ci_exec.sh pip install tabulate
 
       - name: Nightly Test MI35x (1-GPU)
         timeout-minutes: 60
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd-1-gpu-mi35x --nightly --timeout-per-file 600 --continue-on-error || TEST_EXIT_CODE=$?
           echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
@@ -453,20 +453,20 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/amd_ci_install_dependency.sh
+          bash scripts/ci/amd/amd_ci_install_dependency.sh
           # Install tabulate for run_suite.py (missing in MI35x container)
-          bash scripts/ci/amd_ci_exec.sh pip install tabulate
+          bash scripts/ci/amd/amd_ci_exec.sh pip install tabulate
 
       - name: Accuracy Test MI35x (8-GPU GPT-OSS)
         timeout-minutes: 180
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd-8-gpu-mi35x --nightly --timeout-per-file 7200 || TEST_EXIT_CODE=$?
           echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
@@ -485,21 +485,21 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/amd_ci_install_dependency.sh
+          bash scripts/ci/amd/amd_ci_install_dependency.sh
           # Install tabulate for run_suite.py (missing in MI35x container)
-          bash scripts/ci/amd_ci_exec.sh pip install tabulate
+          bash scripts/ci/amd/amd_ci_exec.sh pip install tabulate
 
       - name: Accuracy Test MI35x (8-GPU Grok1-INT4)
         timeout-minutes: 60
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e RCCL_MSCCL_ENABLE=0 \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd-accuracy-8-gpu-mi35x-grok1-int4 --nightly --timeout-per-file 3600 || TEST_EXIT_CODE=$?
@@ -511,7 +511,7 @@ jobs:
         continue-on-error: true  # Perf test failure doesn't fail the job if accuracy passed
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e RCCL_MSCCL_ENABLE=0 \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-perf-8-gpu-mi35x-grok1-int4 --nightly --timeout-per-file 3600 || TEST_EXIT_CODE=$?
@@ -531,21 +531,21 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/amd_ci_install_dependency.sh
+          bash scripts/ci/amd/amd_ci_install_dependency.sh
           # Install tabulate for run_suite.py (missing in MI35x container)
-          bash scripts/ci/amd_ci_exec.sh pip install tabulate
+          bash scripts/ci/amd/amd_ci_exec.sh pip install tabulate
 
       - name: Accuracy Test MI35x (8-GPU Grok2)
         timeout-minutes: 60
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e RCCL_MSCCL_ENABLE=0 \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd-accuracy-8-gpu-mi35x-grok2 --nightly --timeout-per-file 3600 || TEST_EXIT_CODE=$?
@@ -557,7 +557,7 @@ jobs:
         continue-on-error: true  # Perf test failure doesn't fail the job if accuracy passed
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e RCCL_MSCCL_ENABLE=0 \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-perf-8-gpu-mi35x-grok2 --nightly --timeout-per-file 3600 || TEST_EXIT_CODE=$?
@@ -577,21 +577,21 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/amd_ci_install_dependency.sh
+          bash scripts/ci/amd/amd_ci_install_dependency.sh
           # Install tabulate for run_suite.py (missing in MI35x container)
-          bash scripts/ci/amd_ci_exec.sh pip install tabulate
+          bash scripts/ci/amd/amd_ci_exec.sh pip install tabulate
 
       - name: Accuracy Test MI35x (8-GPU DeepSeek-R1-MXFP4)
         timeout-minutes: 180
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd-8-gpu-mi35x-deepseek-r1-mxfp4 --nightly --timeout-per-file 7200 || TEST_EXIT_CODE=$?
           echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
@@ -602,7 +602,7 @@ jobs:
         continue-on-error: true  # Perf test failure doesn't fail the job if accuracy passed
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 registered/amd/perf/mi35x/test_deepseek_r1_mxfp4_perf_mi35x.py || TEST_EXIT_CODE=$?
           echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
@@ -621,21 +621,21 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/amd_ci_install_dependency.sh
+          bash scripts/ci/amd/amd_ci_install_dependency.sh
           # Install tabulate for run_suite.py (missing in MI35x container)
-          bash scripts/ci/amd_ci_exec.sh pip install tabulate
+          bash scripts/ci/amd/amd_ci_exec.sh pip install tabulate
 
       - name: Accuracy Test MI35x (8-GPU DeepSeek-V3.2)
         timeout-minutes: 120
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-amd-8-gpu-mi35x-deepseek-v32 --nightly --timeout-per-file 3600 || TEST_EXIT_CODE=$?
           echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
@@ -654,21 +654,21 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/amd_ci_install_dependency.sh
+          bash scripts/ci/amd/amd_ci_install_dependency.sh
           # Install tabulate for run_suite.py (missing in MI35x container)
-          bash scripts/ci/amd_ci_exec.sh pip install tabulate
+          bash scripts/ci/amd/amd_ci_exec.sh pip install tabulate
 
       - name: Performance Test MI35x (8-GPU DeepSeek-V3.2 Basic)
         timeout-minutes: 150
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-perf-8-gpu-mi35x-deepseek-v32-basic --nightly --timeout-per-file 5400 || TEST_EXIT_CODE=$?
           echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
@@ -687,21 +687,21 @@ jobs:
       - name: Setup docker
         run: |
           touch github_summary.md
-          bash scripts/ci/amd_ci_start_container.sh
+          bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/amd_ci_install_dependency.sh
+          bash scripts/ci/amd/amd_ci_install_dependency.sh
           # Install tabulate for run_suite.py (missing in MI35x container)
-          bash scripts/ci/amd_ci_exec.sh pip install tabulate
+          bash scripts/ci/amd/amd_ci_exec.sh pip install tabulate
 
       - name: Performance Test MI35x (8-GPU DeepSeek-V3.2 MTP)
         timeout-minutes: 150
         run: |
           > github_summary.md  # Clear summary file
-          bash scripts/ci/amd_ci_exec.sh -w /sglang-checkout/test \
+          bash scripts/ci/amd/amd_ci_exec.sh -w /sglang-checkout/test \
             -e GITHUB_STEP_SUMMARY="/sglang-checkout/github_summary.md" \
             python3 run_suite.py --hw amd --suite nightly-perf-8-gpu-mi35x-deepseek-v32-mtp --nightly --timeout-per-file 5400 || TEST_EXIT_CODE=$?
           echo "$(<github_summary.md )" >> $GITHUB_STEP_SUMMARY || true
diff --git a/.github/workflows/nightly-test-npu.yml b/.github/workflows/nightly-test-npu.yml
index b06bf6819..1ab6b6733 100644
--- a/.github/workflows/nightly-test-npu.yml
+++ b/.github/workflows/nightly-test-npu.yml
@@ -51,7 +51,7 @@ jobs:
           pip config set global.extra-index-url "https://pypi.tuna.tsinghua.edu.cn/simple"
           pip config set global.trusted-host "${CACHING_URL} pypi.tuna.tsinghua.edu.cn"
 
-          bash scripts/ci/npu_ci_install_dependency.sh a3
+          bash scripts/ci/npu/npu_ci_install_dependency.sh a3
           # copy required file from our daily cache
           cp ~/.cache/modelscope/hub/datasets/otavia/ShareGPT_Vicuna_unfiltered/ShareGPT_V3_unfiltered_cleaned_split.json /tmp
           # copy download through proxy
@@ -59,7 +59,7 @@ jobs:
 
       - name: Print Log Information
         run: |
-          bash scripts/ci/npu_log_print.sh
+          bash scripts/ci/npu/npu_log_print.sh
       - name: Run test
         timeout-minutes: 240
         env:
@@ -99,7 +99,7 @@ jobs:
           pip config set global.extra-index-url "https://pypi.tuna.tsinghua.edu.cn/simple"
           pip config set global.trusted-host "${CACHING_URL} pypi.tuna.tsinghua.edu.cn"
 
-          bash scripts/ci/npu_ci_install_dependency.sh a3
+          bash scripts/ci/npu/npu_ci_install_dependency.sh a3
           # copy required file from our daily cache
           cp ~/.cache/modelscope/hub/datasets/otavia/ShareGPT_Vicuna_unfiltered/ShareGPT_V3_unfiltered_cleaned_split.json /tmp
           # copy download through proxy
@@ -107,7 +107,7 @@ jobs:
 
       - name: Print Log Information
         run: |
-          bash scripts/ci/npu_log_print.sh
+          bash scripts/ci/npu/npu_log_print.sh
       - name: Run test
         timeout-minutes: 240
         env:
@@ -147,7 +147,7 @@ jobs:
           pip config set global.extra-index-url "https://pypi.tuna.tsinghua.edu.cn/simple"
           pip config set global.trusted-host "${CACHING_URL} pypi.tuna.tsinghua.edu.cn"
 
-          bash scripts/ci/npu_ci_install_dependency.sh a3
+          bash scripts/ci/npu/npu_ci_install_dependency.sh a3
           # copy required file from our daily cache
           cp ~/.cache/modelscope/hub/datasets/otavia/ShareGPT_Vicuna_unfiltered/ShareGPT_V3_unfiltered_cleaned_split.json /tmp
           # copy download through proxy
@@ -155,7 +155,7 @@ jobs:
 
       - name: Print Log Information
         run: |
-          bash scripts/ci/npu_log_print.sh
+          bash scripts/ci/npu/npu_log_print.sh
 
       - name: Run test
         timeout-minutes: 240
diff --git a/.github/workflows/nightly-test-nvidia.yml b/.github/workflows/nightly-test-nvidia.yml
index c97c420fc..d4d84a18d 100644
--- a/.github/workflows/nightly-test-nvidia.yml
+++ b/.github/workflows/nightly-test-nvidia.yml
@@ -66,7 +66,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_dependency.sh
+          bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 60
@@ -86,7 +86,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_dependency.sh
+          bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
@@ -112,7 +112,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_dependency.sh
+          bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run common 8-GPU model tests
         if: always()
@@ -148,7 +148,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_dependency.sh
+          bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
@@ -174,7 +174,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          IS_BLACKWELL=1 bash scripts/ci/ci_install_dependency.sh
+          IS_BLACKWELL=1 bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run common 8-GPU model tests
         if: always()
@@ -199,7 +199,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_dependency.sh
+          bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run eval test for text models
         timeout-minutes: 120
@@ -219,7 +219,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_dependency.sh
+          bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run performance test for text models
         timeout-minutes: 180
@@ -238,7 +238,7 @@ jobs:
           GITHUB_RUN_ID: ${{ github.run_id }}
           GITHUB_RUN_NUMBER: ${{ github.run_number }}
         run: |
-          python3 scripts/ci/publish_traces.py --traces-dir test/performance_profiles_text_models
+          python3 scripts/ci/utils/publish_traces.py --traces-dir test/performance_profiles_text_models
 
   # VLM accuracy tests
   nightly-test-vlm-accuracy-2-gpu-runner:
@@ -252,7 +252,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_dependency.sh
+          bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run eval test for VLM models (fixed MMMU-100)
         timeout-minutes: 240
@@ -272,7 +272,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_dependency.sh
+          bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run perf test for VLM models (MMMU)
         timeout-minutes: 240
@@ -291,7 +291,7 @@ jobs:
           GITHUB_RUN_ID: ${{ github.run_id }}
           GITHUB_RUN_NUMBER: ${{ github.run_number }}
         run: |
-          python3 scripts/ci/publish_traces.py --traces-dir test/performance_profiles_vlms
+          python3 scripts/ci/utils/publish_traces.py --traces-dir test/performance_profiles_vlms
 
   # diffusion performance tests
   nightly-test-multimodal-server-1-gpu:
@@ -310,7 +310,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_dependency.sh diffusion
+          bash scripts/ci/cuda/ci_install_dependency.sh diffusion
           pip install slack_sdk
 
       - name: Run diffusion server tests
@@ -343,7 +343,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_dependency.sh diffusion
+          bash scripts/ci/cuda/ci_install_dependency.sh diffusion
           pip install slack_sdk
 
       - name: Run diffusion server tests
@@ -371,7 +371,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          IS_BLACKWELL=1 bash scripts/ci/ci_install_dependency.sh
+          IS_BLACKWELL=1 bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 300
@@ -393,7 +393,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          IS_BLACKWELL=1 bash scripts/ci/ci_install_dependency.sh
+          IS_BLACKWELL=1 bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 120
diff --git a/.github/workflows/pr-benchmark-rust.yml b/.github/workflows/pr-benchmark-rust.yml
index 85f174b41..bbee6bc71 100644
--- a/.github/workflows/pr-benchmark-rust.yml
+++ b/.github/workflows/pr-benchmark-rust.yml
@@ -34,7 +34,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_gateway_dependencies.sh
+          bash scripts/ci/cuda/ci_install_gateway_dependencies.sh
 
       - name: Configure sccache
         uses: mozilla-actions/sccache-action@v0.0.9
@@ -117,7 +117,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_gateway_dependencies.sh
+          bash scripts/ci/cuda/ci_install_gateway_dependencies.sh
 
       - name: Configure sccache
         uses: mozilla-actions/sccache-action@v0.0.9
diff --git a/.github/workflows/pr-test-amd.yml b/.github/workflows/pr-test-amd.yml
index 5c05d9640..1c3a3cc1c 100644
--- a/.github/workflows/pr-test-amd.yml
+++ b/.github/workflows/pr-test-amd.yml
@@ -92,7 +92,8 @@ jobs:
               - "python/sglang/!(multimodal_gen)/**"
               - "python/pyproject_rocm.toml"
               - "python/pyproject_other.toml"
-              - "scripts/ci/**"
+              - "scripts/ci/amd/*"
+              - "scripts/ci/utils/*"
               - "test/**"
               - ".github/workflows/pr-test-amd.yml"
             sgl_kernel:
@@ -131,13 +132,13 @@ jobs:
         run: bash scripts/ensure_vram_clear.sh rocm
 
       - name: Start CI container
-        run: bash scripts/ci/amd_ci_start_container.sh
+        run: bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/amd_ci_install_dependency.sh
+          bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 14
@@ -179,13 +180,13 @@ jobs:
         run: bash scripts/ensure_vram_clear.sh rocm
 
       - name: Start CI container
-        run: bash scripts/ci/amd_ci_start_container.sh
+        run: bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/amd_ci_install_dependency.sh
+          bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 20
@@ -222,18 +223,18 @@ jobs:
         run: bash scripts/ensure_vram_clear.sh rocm
 
       - name: Start CI container
-        run: bash scripts/ci/amd_ci_start_container.sh
+        run: bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/amd_ci_install_dependency.sh
+          bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 10
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-a-test-1-amd
+          bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-a-test-1-amd
 
   stage-b-test-small-1-gpu-amd:
     needs: [check-changes, stage-a-test-1-amd]
@@ -263,17 +264,17 @@ jobs:
         run: bash scripts/ensure_vram_clear.sh rocm
 
       - name: Start CI container
-        run: bash scripts/ci/amd_ci_start_container.sh
+        run: bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-small-1-gpu-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 13 --timeout-per-file 1800
+          bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-small-1-gpu-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 13 --timeout-per-file 1800
 
   stage-b-test-small-1-gpu-amd-mi35x:
     needs: [check-changes, stage-a-test-1-amd]
@@ -302,17 +303,17 @@ jobs:
         run: bash scripts/ensure_vram_clear.sh rocm
 
       - name: Start CI container
-        run: bash scripts/ci/amd_ci_start_container.sh
+        run: bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-small-1-gpu-amd-mi35x
+          bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-small-1-gpu-amd-mi35x
 
   stage-b-test-large-2-gpu-amd:
     needs: [check-changes, stage-a-test-1-amd]
@@ -341,17 +342,17 @@ jobs:
         run: bash scripts/ensure_vram_clear.sh rocm
 
       - name: Start CI container
-        run: bash scripts/ci/amd_ci_start_container.sh
+        run: bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-large-2-gpu-amd
+          bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-large-2-gpu-amd
 
   multimodal-gen-test-1-gpu-amd:
     needs: [check-changes]
@@ -381,13 +382,13 @@ jobs:
           pattern: wheel-python3.10-cuda12.9
 
       - name: Start CI container
-        run: bash scripts/ci/amd_ci_start_container.sh
+        run: bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/amd_ci_install_dependency.sh diffusion
+          bash scripts/ci/amd/amd_ci_install_dependency.sh diffusion
 
       - name: Setup kernel caches
         run: |
@@ -501,13 +502,13 @@ jobs:
           pattern: wheel-python3.10-cuda12.9
 
       - name: Start CI container
-        run: bash scripts/ci/amd_ci_start_container.sh
+        run: bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/amd_ci_install_dependency.sh diffusion
+          bash scripts/ci/amd/amd_ci_install_dependency.sh diffusion
 
       - name: Setup kernel caches
         run: |
@@ -622,23 +623,23 @@ jobs:
         run: bash scripts/ensure_vram_clear.sh rocm
 
       - name: Start CI container
-        run: bash scripts/ci/amd_ci_start_container.sh
+        run: bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Test RCCL multi-GPU communication
         timeout-minutes: 5
         run: |
           echo "Testing RCCL multi-GPU communication with debug info..."
-          docker exec ci_sglang bash -c "cd /sglang-checkout && NCCL_DEBUG=INFO RCCL_DEBUG=INFO torchrun --nproc_per_node=8 scripts/ci/test_rccl_multi_gpu.py"
+          docker exec ci_sglang bash -c "cd /sglang-checkout && NCCL_DEBUG=INFO RCCL_DEBUG=INFO torchrun --nproc_per_node=8 scripts/ci/amd/test_rccl_multi_gpu.py"
 
       - name: Run test
         timeout-minutes: 60
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-c-test-large-8-gpu-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 2 --timeout-per-file 3600
+          bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-c-test-large-8-gpu-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 2 --timeout-per-file 3600
 
   stage-c-test-large-8-gpu-amd-mi35x:
     needs: [check-changes, call-gate, stage-b-test-small-1-gpu-amd, stage-b-test-large-2-gpu-amd]
@@ -668,17 +669,17 @@ jobs:
         run: bash scripts/ensure_vram_clear.sh rocm
 
       - name: Start CI container
-        run: bash scripts/ci/amd_ci_start_container.sh
+        run: bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 60
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-c-test-large-8-gpu-amd-mi35x --auto-partition-id ${{ matrix.part }} --auto-partition-size 2 --timeout-per-file 3600
+          bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-c-test-large-8-gpu-amd-mi35x --auto-partition-id ${{ matrix.part }} --auto-partition-size 2 --timeout-per-file 3600
 
   stage-b-test-small-1-gpu-performance-amd:
     needs: [check-changes, call-gate, stage-a-test-1-amd]
@@ -707,17 +708,17 @@ jobs:
         run: bash scripts/ensure_vram_clear.sh rocm
 
       - name: Start CI container
-        run: bash scripts/ci/amd_ci_start_container.sh
+        run: bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-small-1-gpu-performance-amd --timeout-per-file 1200
+          bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-small-1-gpu-performance-amd --timeout-per-file 1200
 
   stage-b-test-large-1-gpu-performance-amd:
     needs: [check-changes, call-gate, stage-a-test-1-amd]
@@ -747,17 +748,17 @@ jobs:
         run: bash scripts/ensure_vram_clear.sh rocm
 
       - name: Start CI container
-        run: bash scripts/ci/amd_ci_start_container.sh
+        run: bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-large-1-gpu-performance-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 2 --timeout-per-file 1200
+          bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-large-1-gpu-performance-amd --auto-partition-id ${{ matrix.part }} --auto-partition-size 2 --timeout-per-file 1200
 
   stage-b-test-large-2-gpu-performance-amd:
     needs: [check-changes, call-gate, stage-a-test-1-amd]
@@ -786,17 +787,17 @@ jobs:
         run: bash scripts/ensure_vram_clear.sh rocm
 
       - name: Start CI container
-        run: bash scripts/ci/amd_ci_start_container.sh
+        run: bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-large-2-gpu-performance-amd --timeout-per-file 1200
+          bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" python3 run_suite.py --hw amd --suite stage-b-test-large-2-gpu-performance-amd --timeout-per-file 1200
 
   stage-b-test-small-1-gpu-accuracy-amd:
     needs: [check-changes, stage-a-test-1-amd]
@@ -825,17 +826,17 @@ jobs:
         run: bash scripts/ensure_vram_clear.sh rocm
 
       - name: Start CI container
-        run: bash scripts/ci/amd_ci_start_container.sh
+        run: bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w "/sglang-checkout/test" -e SGLANG_USE_AITER=0 python3 run_suite.py --hw amd --suite stage-b-test-small-1-gpu-accuracy-amd --timeout-per-file 1800
+          bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" -e SGLANG_USE_AITER=0 python3 run_suite.py --hw amd --suite stage-b-test-small-1-gpu-accuracy-amd --timeout-per-file 1800
 
   stage-b-test-large-2-gpu-accuracy-amd:
     needs: [check-changes, stage-a-test-1-amd]
@@ -864,17 +865,17 @@ jobs:
         run: bash scripts/ensure_vram_clear.sh rocm
 
       - name: Start CI container
-        run: bash scripts/ci/amd_ci_start_container.sh
+        run: bash scripts/ci/amd/amd_ci_start_container.sh
         env:
           GITHUB_WORKSPACE: ${{ github.workspace }}
 
       - name: Install dependencies
-        run: bash scripts/ci/amd_ci_install_dependency.sh
+        run: bash scripts/ci/amd/amd_ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
         run: |
-          bash scripts/ci/amd_ci_exec.sh -w "/sglang-checkout/test" -e SGLANG_USE_AITER_AR=0 -e SGLANG_USE_AITER=0 -e HF_HUB_ENABLE_HF_TRANSFER=0 python3 run_suite.py --hw amd --suite stage-b-test-large-2-gpu-accuracy-amd --timeout-per-file 1800
+          bash scripts/ci/amd/amd_ci_exec.sh -w "/sglang-checkout/test" -e SGLANG_USE_AITER_AR=0 -e SGLANG_USE_AITER=0 -e HF_HUB_ENABLE_HF_TRANSFER=0 python3 run_suite.py --hw amd --suite stage-b-test-large-2-gpu-accuracy-amd --timeout-per-file 1800
 
   pr-test-amd-finish:
     needs:
diff --git a/.github/workflows/pr-test-npu.yml b/.github/workflows/pr-test-npu.yml
index e908538ab..644464167 100644
--- a/.github/workflows/pr-test-npu.yml
+++ b/.github/workflows/pr-test-npu.yml
@@ -57,7 +57,7 @@ jobs:
             main_package:
               - "python/sglang/!(multimodal_gen)/**"
               - "python/pyproject_other.toml"
-              - "scripts/ci/npu_ci_install_dependency.sh"
+              - "scripts/ci/npu/npu_ci_install_dependency.sh"
               - "test/srt/ascend/**"
               - ".github/workflows/pr-test-npu.yml"
 
@@ -89,7 +89,7 @@ jobs:
           pip config set global.extra-index-url "https://pypi.tuna.tsinghua.edu.cn/simple"
           pip config set global.trusted-host "${CACHING_URL} pypi.tuna.tsinghua.edu.cn"
 
-          bash scripts/ci/npu_ci_install_dependency.sh 910b
+          bash scripts/ci/npu/npu_ci_install_dependency.sh 910b
           # copy required file from our daily cache
           cp ~/.cache/modelscope/hub/datasets/otavia/ShareGPT_Vicuna_unfiltered/ShareGPT_V3_unfiltered_cleaned_split.json /tmp
           # copy download through proxy
@@ -134,7 +134,7 @@ jobs:
           pip config set global.extra-index-url "https://pypi.tuna.tsinghua.edu.cn/simple"
           pip config set global.trusted-host "${CACHING_URL} pypi.tuna.tsinghua.edu.cn"
 
-          bash scripts/ci/npu_ci_install_dependency.sh 910b
+          bash scripts/ci/npu/npu_ci_install_dependency.sh 910b
           # copy required file from our daily cache
           cp ~/.cache/modelscope/hub/datasets/otavia/ShareGPT_Vicuna_unfiltered/ShareGPT_V3_unfiltered_cleaned_split.json /tmp
           # copy download through proxy
@@ -175,7 +175,7 @@ jobs:
           pip config set global.extra-index-url "https://pypi.tuna.tsinghua.edu.cn/simple"
           pip config set global.trusted-host "${CACHING_URL} pypi.tuna.tsinghua.edu.cn"
 
-          bash scripts/ci/npu_ci_install_dependency.sh 910b
+          bash scripts/ci/npu/npu_ci_install_dependency.sh 910b
           # copy required file from our daily cache
           cp ~/.cache/modelscope/hub/datasets/otavia/ShareGPT_Vicuna_unfiltered/ShareGPT_V3_unfiltered_cleaned_split.json /tmp
           # copy download through proxy
@@ -220,7 +220,7 @@ jobs:
           pip config set global.extra-index-url "https://pypi.tuna.tsinghua.edu.cn/simple"
           pip config set global.trusted-host "${CACHING_URL} pypi.tuna.tsinghua.edu.cn"
 
-          bash scripts/ci/npu_ci_install_dependency.sh a3
+          bash scripts/ci/npu/npu_ci_install_dependency.sh a3
           # copy required file from our daily cache
           cp ~/.cache/modelscope/hub/datasets/otavia/ShareGPT_Vicuna_unfiltered/ShareGPT_V3_unfiltered_cleaned_split.json /tmp
           # copy download through proxy
diff --git a/.github/workflows/pr-test-rust.yml b/.github/workflows/pr-test-rust.yml
index 9b707f4f7..4202f6c2e 100644
--- a/.github/workflows/pr-test-rust.yml
+++ b/.github/workflows/pr-test-rust.yml
@@ -33,7 +33,7 @@ jobs:
 
       - name: Install rust dependencies
         run: |
-          bash scripts/ci/ci_install_gateway_dependencies.sh
+          bash scripts/ci/cuda/ci_install_gateway_dependencies.sh
 
       - name: Configure sccache
         uses: mozilla-actions/sccache-action@v0.0.9
@@ -120,7 +120,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_gateway_dependencies.sh
+          bash scripts/ci/cuda/ci_install_gateway_dependencies.sh
 
       - name: Configure sccache
         uses: mozilla-actions/sccache-action@v0.0.9
@@ -222,7 +222,7 @@ jobs:
 
       - name: Install SGLang dependencies
         run: |
-          sudo --preserve-env=PATH bash scripts/ci/ci_install_dependency.sh
+          sudo --preserve-env=PATH bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Setup Oracle Instant Client
         if: matrix.setup_oracle
diff --git a/.github/workflows/pr-test-xeon.yml b/.github/workflows/pr-test-xeon.yml
index c56a281f4..021a13085 100644
--- a/.github/workflows/pr-test-xeon.yml
+++ b/.github/workflows/pr-test-xeon.yml
@@ -57,7 +57,6 @@ jobs:
             main_package:
               - "python/sglang/!(multimodal_gen)/**"
               - "python/pyproject_cpu.toml"
-              - "scripts/ci/**"
               - "test/**"
               - "sgl-kernel/**"
               - ".github/workflows/pr-test-xeon.yml"
diff --git a/.github/workflows/pr-test-xpu.yml b/.github/workflows/pr-test-xpu.yml
index 5785eb2ef..38b89762a 100644
--- a/.github/workflows/pr-test-xpu.yml
+++ b/.github/workflows/pr-test-xpu.yml
@@ -56,7 +56,6 @@ jobs:
             main_package:
               - "python/sglang/!(multimodal_gen)/**"
               - "python/pyproject_xpu.toml"
-              - "scripts/ci/**"
               - "test/**"
               - "sgl-kernel/**"
               - ".github/workflows/pr-test-xpu.yml"
diff --git a/.github/workflows/pr-test.yml b/.github/workflows/pr-test.yml
index 820357dde..8e54154a2 100644
--- a/.github/workflows/pr-test.yml
+++ b/.github/workflows/pr-test.yml
@@ -108,7 +108,8 @@ jobs:
             main_package:
               - "python/sglang/!(multimodal_gen)/**"
               - "python/pyproject.toml"
-              - "scripts/ci/**"
+              - "scripts/ci/cuda/*"
+              - "scripts/ci/utils/*"
               - "test/**"
               - ".github/workflows/pr-test.yml"
             sgl_kernel:
@@ -502,7 +503,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh diffusion
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh diffusion
 
       - name: Run test
         timeout-minutes: 30
@@ -542,7 +543,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
@@ -583,7 +584,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run benchmark tests
         timeout-minutes: 45
@@ -635,7 +636,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} IS_BLACKWELL=1 bash scripts/ci/ci_install_dependency.sh diffusion
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} IS_BLACKWELL=1 bash scripts/ci/cuda/ci_install_dependency.sh diffusion
 
       - name: Run sgl-kernel unit tests on B200
         timeout-minutes: 30
@@ -666,7 +667,7 @@ jobs:
 
   #     - name: Install dependencies
   #       run: |
-  #         CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+  #         CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
 
   #     - name: Run kernel unit tests
   #       timeout-minutes: 30
@@ -696,7 +697,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          bash scripts/ci/ci_install_dependency.sh
+          bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
@@ -739,7 +740,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 10
@@ -839,7 +840,7 @@ jobs:
         timeout-minutes: 10
         run: |
           source /etc/profile.d/sglang-ci.sh
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
@@ -891,7 +892,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
@@ -940,7 +941,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
@@ -987,7 +988,7 @@ jobs:
         timeout-minutes: 10
         run: |
           source /etc/profile.d/sglang-ci.sh
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
@@ -1037,7 +1038,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 40
@@ -1082,7 +1083,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
@@ -1129,7 +1130,7 @@ jobs:
         timeout-minutes: 10
         run: |
           source /etc/profile.d/sglang-ci.sh
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
           git clone https://github.com/merrymercy/human-eval.git
           cd human-eval
           pip install -e .
@@ -1178,7 +1179,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
           git clone https://github.com/merrymercy/human-eval.git
           cd human-eval
           pip install -e .
@@ -1229,7 +1230,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} IS_BLACKWELL=1 bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} IS_BLACKWELL=1 bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
@@ -1279,7 +1280,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
@@ -1324,7 +1325,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} IS_BLACKWELL=1 bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} IS_BLACKWELL=1 bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
@@ -1367,7 +1368,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh diffusion
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh diffusion
       - name: Run diffusion server tests
         timeout-minutes: 240
         run: |
@@ -1418,7 +1419,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh diffusion
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh diffusion
 
       - name: Run diffusion server tests
         timeout-minutes: 240
@@ -1471,7 +1472,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 20
@@ -1524,7 +1525,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_dependency.sh
 
       # - name: Warmup Weights and JIT Compilation
       #   timeout-minutes: 20
@@ -1584,7 +1585,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_deepep.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_deepep.sh
 
       - name: Run test
         timeout-minutes: 20
@@ -1633,7 +1634,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_deepep.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_deepep.sh
 
       - name: Run test
         timeout-minutes: 20
@@ -1682,7 +1683,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_deepep.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/cuda/ci_install_deepep.sh
 
       - name: Run test
         timeout-minutes: 45
@@ -1736,7 +1737,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} IS_BLACKWELL=1 bash scripts/ci/ci_install_dependency.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} IS_BLACKWELL=1 bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 30
@@ -1787,7 +1788,7 @@ jobs:
       - name: Install dependencies
         timeout-minutes: 10
         run: |
-          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} IS_BLACKWELL=1 GRACE_BLACKWELL=1 bash scripts/ci/ci_install_deepep.sh
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} IS_BLACKWELL=1 GRACE_BLACKWELL=1 bash scripts/ci/cuda/ci_install_deepep.sh
 
       - name: Run test
         timeout-minutes: 45
diff --git a/.github/workflows/release-docs.yml b/.github/workflows/release-docs.yml
index 78fafc60b..07e6871a2 100644
--- a/.github/workflows/release-docs.yml
+++ b/.github/workflows/release-docs.yml
@@ -24,7 +24,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_dependency.sh
+          bash scripts/ci/cuda/ci_install_dependency.sh
           pip install -r docs/requirements.txt
           apt-get update && apt-get install -y pandoc parallel retry
           ln -sf "$(which python3)" /usr/bin/python
diff --git a/.github/workflows/runner-utilization.yml b/.github/workflows/runner-utilization.yml
index b56cd1fe5..7c37e41de 100644
--- a/.github/workflows/runner-utilization.yml
+++ b/.github/workflows/runner-utilization.yml
@@ -6,7 +6,7 @@ on:
   pull_request:
     paths:
       - '.github/workflows/runner-utilization.yml'
-      - 'scripts/ci/runner_utilization_report.py'
+      - 'scripts/ci/utils/runner_utilization_report.py'
   workflow_dispatch:
     inputs:
       hours:
@@ -37,7 +37,7 @@ jobs:
         env:
           GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
         run: |
-          python scripts/ci/runner_utilization_report.py \
+          python scripts/ci/utils/runner_utilization_report.py \
             --repo ${{ github.repository }} \
             --hours ${{ inputs.hours || '24' }} \
             ${{ inputs.filter && format('--filter {0}', inputs.filter) || '' }}
diff --git a/.github/workflows/slash-command-handler.yml b/.github/workflows/slash-command-handler.yml
index 1293dc845..012208f9f 100644
--- a/.github/workflows/slash-command-handler.yml
+++ b/.github/workflows/slash-command-handler.yml
@@ -75,4 +75,4 @@ jobs:
           COMMENT_BODY: ${{ github.event.comment.body }}
           USER_LOGIN: ${{ github.event.comment.user.login }}
         run: |
-          python scripts/ci/slash_command_handler.py
+          python scripts/ci/utils/slash_command_handler.py
diff --git a/.github/workflows/stress-test.yml b/.github/workflows/stress-test.yml
index 4e6bc741d..241c097e8 100644
--- a/.github/workflows/stress-test.yml
+++ b/.github/workflows/stress-test.yml
@@ -24,7 +24,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_dependency.sh
+          bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run stress tests
         timeout-minutes: 210
diff --git a/.github/workflows/weekly-test-nvidia.yml b/.github/workflows/weekly-test-nvidia.yml
index 0a05cdbc3..01769cf02 100644
--- a/.github/workflows/weekly-test-nvidia.yml
+++ b/.github/workflows/weekly-test-nvidia.yml
@@ -37,7 +37,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          bash scripts/ci/ci_install_dependency.sh
+          bash scripts/ci/cuda/ci_install_dependency.sh
 
       - name: Run weekly 8-GPU H200 tests
         timeout-minutes: 120
diff --git a/scripts/ci/amd_ci_exec.sh b/scripts/ci/amd/amd_ci_exec.sh
similarity index 100%
rename from scripts/ci/amd_ci_exec.sh
rename to scripts/ci/amd/amd_ci_exec.sh
diff --git a/scripts/ci/amd_ci_install_dependency.sh b/scripts/ci/amd/amd_ci_install_dependency.sh
similarity index 98%
rename from scripts/ci/amd_ci_install_dependency.sh
rename to scripts/ci/amd/amd_ci_install_dependency.sh
index ac0303881..271adfaa7 100755
--- a/scripts/ci/amd_ci_install_dependency.sh
+++ b/scripts/ci/amd/amd_ci_install_dependency.sh
@@ -252,4 +252,4 @@ docker exec ci_sglang ls -la /sgl-workspace/aiter/aiter/jit/ 2>/dev/null || echo
 
 # Pre-build AITER kernels to avoid timeout during tests
 echo "Warming up AITER JIT kernels..."
-docker exec -e SGLANG_USE_AITER=1 ci_sglang python3 /sglang-checkout/scripts/ci/amd_ci_warmup_aiter.py || echo "AITER warmup completed (some kernels may not be available)"
+docker exec -e SGLANG_USE_AITER=1 ci_sglang python3 /sglang-checkout/scripts/ci/amd/amd_ci_warmup_aiter.py || echo "AITER warmup completed (some kernels may not be available)"
diff --git a/scripts/ci/amd_ci_start_container.sh b/scripts/ci/amd/amd_ci_start_container.sh
similarity index 100%
rename from scripts/ci/amd_ci_start_container.sh
rename to scripts/ci/amd/amd_ci_start_container.sh
diff --git a/scripts/ci/amd_ci_warmup_aiter.py b/scripts/ci/amd/amd_ci_warmup_aiter.py
similarity index 100%
rename from scripts/ci/amd_ci_warmup_aiter.py
rename to scripts/ci/amd/amd_ci_warmup_aiter.py
diff --git a/scripts/ci/test_rccl_multi_gpu.py b/scripts/ci/amd/test_rccl_multi_gpu.py
similarity index 100%
rename from scripts/ci/test_rccl_multi_gpu.py
rename to scripts/ci/amd/test_rccl_multi_gpu.py
diff --git a/scripts/ci/ci_install_deepep.sh b/scripts/ci/cuda/ci_install_deepep.sh
similarity index 98%
rename from scripts/ci/ci_install_deepep.sh
rename to scripts/ci/cuda/ci_install_deepep.sh
index 92258c19f..9e54df24c 100755
--- a/scripts/ci/ci_install_deepep.sh
+++ b/scripts/ci/cuda/ci_install_deepep.sh
@@ -2,7 +2,7 @@
 # Install the dependency in CI.
 set -euxo pipefail
 
-bash scripts/ci/ci_install_dependency.sh
+bash scripts/ci/cuda/ci_install_dependency.sh
 
 export GDRCOPY_HOME=/usr/src/gdrdrv-2.5.1/
 export CUDA_HOME=/usr/local/cuda
diff --git a/scripts/ci/ci_install_dependency.sh b/scripts/ci/cuda/ci_install_dependency.sh
similarity index 99%
rename from scripts/ci/ci_install_dependency.sh
rename to scripts/ci/cuda/ci_install_dependency.sh
index f9ef787c9..f88ec7f89 100755
--- a/scripts/ci/ci_install_dependency.sh
+++ b/scripts/ci/cuda/ci_install_dependency.sh
@@ -20,7 +20,7 @@ fi
 
 # Kill existing processes
 SCRIPT_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"
-bash "${SCRIPT_DIR}/../killall_sglang.sh"
+bash "${SCRIPT_DIR}/../../killall_sglang.sh"
 echo "CUDA_VISIBLE_DEVICES=${CUDA_VISIBLE_DEVICES:-}"
 
 # Clear torch compilation cache
diff --git a/scripts/ci/ci_install_gateway_dependencies.sh b/scripts/ci/cuda/ci_install_gateway_dependencies.sh
similarity index 100%
rename from scripts/ci/ci_install_gateway_dependencies.sh
rename to scripts/ci/cuda/ci_install_gateway_dependencies.sh
diff --git a/scripts/ci/ci_start_disaggregation_servers.sh b/scripts/ci/cuda/ci_start_disaggregation_servers.sh
similarity index 100%
rename from scripts/ci/ci_start_disaggregation_servers.sh
rename to scripts/ci/cuda/ci_start_disaggregation_servers.sh
diff --git a/scripts/ci/prepare_runner.sh b/scripts/ci/cuda/prepare_runner.sh
similarity index 81%
rename from scripts/ci/prepare_runner.sh
rename to scripts/ci/cuda/prepare_runner.sh
index 3a6701bf2..2ef84cbda 100755
--- a/scripts/ci/prepare_runner.sh
+++ b/scripts/ci/cuda/prepare_runner.sh
@@ -8,12 +8,12 @@ echo "Preparing CI runner..."
 echo ""
 
 # Clean up stale HuggingFace cache artifacts from previous failed downloads
-python3 "${SCRIPT_DIR}/cleanup_hf_cache.py"
+python3 "${SCRIPT_DIR}/../utils/cleanup_hf_cache.py"
 echo ""
 
 # Pre-validate cached models and write markers for offline mode
 # This allows tests to run with HF_HUB_OFFLINE=1 for models that are fully cached
-python3 "${SCRIPT_DIR}/prevalidate_cached_models.py"
+python3 "${SCRIPT_DIR}/../utils/prevalidate_cached_models.py"
 echo ""
 
 echo "CI runner preparation complete!"
diff --git a/scripts/ci/npu_ci_install_dependency.sh b/scripts/ci/npu/npu_ci_install_dependency.sh
similarity index 100%
rename from scripts/ci/npu_ci_install_dependency.sh
rename to scripts/ci/npu/npu_ci_install_dependency.sh
diff --git a/scripts/ci/npu_log_print.sh b/scripts/ci/npu/npu_log_print.sh
similarity index 96%
rename from scripts/ci/npu_log_print.sh
rename to scripts/ci/npu/npu_log_print.sh
index ad59bc594..92ba4fe3e 100755
--- a/scripts/ci/npu_log_print.sh
+++ b/scripts/ci/npu/npu_log_print.sh
@@ -23,4 +23,4 @@ get_sha() {
 }
 get_sha "sglang" "$SGLANG_URL" "$SGLANG_BRANCH"
 get_sha "sgl-kernel" "$SGL_KERNEL_URL" "$SGL_KERNEL_BRANCH"
-chmod +x scripts/ci/npu_log_print.sh
+chmod +x scripts/ci/npu/npu_log_print.sh
diff --git a/scripts/ci/ci_coverage_report.py b/scripts/ci/utils/ci_coverage_report.py
similarity index 98%
rename from scripts/ci/ci_coverage_report.py
rename to scripts/ci/utils/ci_coverage_report.py
index 2ba07d560..1dc6708d6 100755
--- a/scripts/ci/ci_coverage_report.py
+++ b/scripts/ci/utils/ci_coverage_report.py
@@ -6,7 +6,7 @@ Collects all CI test registrations from test/registered/ and generates
 a coverage report organized by folder, backend, and suite.
 
 Usage:
-    python scripts/ci/ci_coverage_report.py [--output-format markdown|json]
+    python scripts/ci/utils/ci_coverage_report.py [--output-format markdown|json]
 """
 
 import argparse
@@ -19,7 +19,10 @@ from pathlib import Path
 
 # Add the ci_register module path directly to avoid heavy sglang imports
 sys.path.insert(
-    0, str(Path(__file__).parent.parent.parent / "python" / "sglang" / "test" / "ci")
+    0,
+    str(
+        Path(__file__).parent.parent.parent.parent / "python" / "sglang" / "test" / "ci"
+    ),
 )
 
 from ci_register import CIRegistry, HWBackend, ut_parse_one_file
@@ -450,7 +453,7 @@ def main():
     args = parser.parse_args()
 
     # Change to repo root if needed
-    script_dir = Path(__file__).parent
+    script_dir = Path(__file__).parent.parent
     repo_root = script_dir.parent.parent
     os.chdir(repo_root)
 
diff --git a/scripts/ci/cleanup_hf_cache.py b/scripts/ci/utils/cleanup_hf_cache.py
similarity index 100%
rename from scripts/ci/cleanup_hf_cache.py
rename to scripts/ci/utils/cleanup_hf_cache.py
diff --git a/scripts/ci/prevalidate_cached_models.py b/scripts/ci/utils/prevalidate_cached_models.py
similarity index 99%
rename from scripts/ci/prevalidate_cached_models.py
rename to scripts/ci/utils/prevalidate_cached_models.py
index 9d08f987b..9227eb109 100755
--- a/scripts/ci/prevalidate_cached_models.py
+++ b/scripts/ci/utils/prevalidate_cached_models.py
@@ -20,7 +20,7 @@ import time
 from pathlib import Path
 
 # Add python directory to path to import sglang modules
-REPO_ROOT = Path(__file__).parent.parent.parent
+REPO_ROOT = Path(__file__).parent.parent.parent.parent
 sys.path.insert(0, str(REPO_ROOT / "python"))
 
 from sglang.srt.model_loader.ci_weight_validation import (  # noqa: E402
diff --git a/scripts/ci/publish_traces.py b/scripts/ci/utils/publish_traces.py
similarity index 100%
rename from scripts/ci/publish_traces.py
rename to scripts/ci/utils/publish_traces.py
diff --git a/scripts/ci/runner_utilization_report.py b/scripts/ci/utils/runner_utilization_report.py
similarity index 100%
rename from scripts/ci/runner_utilization_report.py
rename to scripts/ci/utils/runner_utilization_report.py
diff --git a/scripts/ci/slash_command_handler.py b/scripts/ci/utils/slash_command_handler.py
similarity index 100%
rename from scripts/ci/slash_command_handler.py
rename to scripts/ci/utils/slash_command_handler.py
diff --git a/scripts/release/utils.py b/scripts/release/utils.py
index da212734b..62817a70c 100644
--- a/scripts/release/utils.py
+++ b/scripts/release/utils.py
@@ -80,7 +80,7 @@ def compare_versions(v1: str, v2: str) -> int:
 
 
 def get_repo_root() -> Path:
-    return Path(__file__).parent.parent.parent
+    return Path(__file__).parent.parent.parent.parent
 
 
 def read_current_version(version_file: Path) -> str: