diff --git a/.buildkite/test-amd.yaml b/.buildkite/test-amd.yaml
index 0b06aab230..fe4e4fd9f2 100644
--- a/.buildkite/test-amd.yaml
+++ b/.buildkite/test-amd.yaml
@@ -3869,6 +3869,84 @@ steps:
   - uv pip install --system 'evalscope==1.10.0'
   - pytest -s -v evals/qwen4_exp/test_accuracy.py --config-list-file=configs/models-rocm.txt
 
+- label: ":amd: (MI355) LM Eval TurboQuant k8v4"
+  timeout_in_minutes: 30
+  mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
+  dind: false
+  agent_pool: mi355_1
+  num_gpus: 1
+  optional: true
+  working_dir: "/vllm-workspace/tests"
+  source_file_dependencies:
+  - tests/evals/gsm8k/
+  - tests/utils.py
+  - tests/conftest.py
+  - vllm/model_executor/layers/quantization/turboquant/
+  - vllm/v1/attention/backends/turboquant_attn.py
+  - vllm/v1/attention/backends/triton_attn.py
+  - vllm/v1/attention/ops/triton_unified_attention.py
+  - vllm/v1/attention/selector.py
+  - vllm/v1/attention/ops/triton_turboquant_decode.py
+  - vllm/v1/attention/ops/triton_turboquant_store.py
+  - vllm/v1/attention/ops/flydsl_turboquant_decode.py
+  - vllm/v1/attention/ops/flydsl_kernels/
+  - vllm/v1/attention/ops/turboquant_soa/
+  - vllm/platforms/rocm.py
+  commands:
+  - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=evals/gsm8k/configs/models-turboquant-k8v4-rocm.txt
+
+- label: ":amd: (MI355) LM Eval TurboQuant t4nc"
+  timeout_in_minutes: 30
+  mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
+  dind: false
+  agent_pool: mi355_1
+  num_gpus: 1
+  optional: true
+  working_dir: "/vllm-workspace/tests"
+  source_file_dependencies:
+  - tests/evals/gsm8k/
+  - tests/utils.py
+  - tests/conftest.py
+  - vllm/model_executor/layers/quantization/turboquant/
+  - vllm/v1/attention/backends/turboquant_attn.py
+  - vllm/v1/attention/backends/triton_attn.py
+  - vllm/v1/attention/ops/triton_unified_attention.py
+  - vllm/v1/attention/selector.py
+  - vllm/v1/attention/ops/triton_turboquant_decode.py
+  - vllm/v1/attention/ops/triton_turboquant_store.py
+  - vllm/v1/attention/ops/flydsl_turboquant_decode.py
+  - vllm/v1/attention/ops/flydsl_kernels/
+  - vllm/v1/attention/ops/turboquant_soa/
+  - vllm/platforms/rocm.py
+  commands:
+  - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=evals/gsm8k/configs/models-turboquant-t4nc-rocm.txt
+
+- label: ":amd: (MI355) LM Eval TurboQuant k3v4nc"
+  timeout_in_minutes: 30
+  mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
+  dind: false
+  agent_pool: mi355_1
+  num_gpus: 1
+  optional: true
+  working_dir: "/vllm-workspace/tests"
+  source_file_dependencies:
+  - tests/evals/gsm8k/
+  - tests/utils.py
+  - tests/conftest.py
+  - vllm/model_executor/layers/quantization/turboquant/
+  - vllm/v1/attention/backends/turboquant_attn.py
+  - vllm/v1/attention/backends/triton_attn.py
+  - vllm/v1/attention/ops/triton_unified_attention.py
+  - vllm/v1/attention/selector.py
+  - vllm/v1/attention/ops/triton_turboquant_decode.py
+  - vllm/v1/attention/ops/triton_turboquant_store.py
+  - vllm/v1/attention/ops/flydsl_turboquant_decode.py
+  - vllm/v1/attention/ops/flydsl_kernels/
+  - vllm/v1/attention/ops/turboquant_soa/
+  - vllm/platforms/rocm.py
+  commands:
+  - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=evals/gsm8k/configs/models-turboquant-k3v4nc-rocm.txt
+
 - label: ":amd: (MI355) LM Eval Small Models Distributed"
   timeout_in_minutes: 55
   mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
diff --git a/.buildkite/test_areas/lm_eval.yaml b/.buildkite/test_areas/lm_eval.yaml
index eaa8218047..f1ef63b4ed 100644
--- a/.buildkite/test_areas/lm_eval.yaml
+++ b/.buildkite/test_areas/lm_eval.yaml
@@ -646,6 +646,34 @@ steps:
   - vllm/v1/attention/ops/triton_turboquant_store.py
   commands:
   - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=evals/gsm8k/configs/models-turboquant-k8v4.txt
+  mirror:
+    amd:
+      label: ":amd: (MI355) LM Eval TurboQuant k8v4"
+      dind: false
+      device: mi355_1
+      num_devices: 1
+      timeout_in_minutes: 30
+      working_dir: /vllm-workspace/tests
+      depends_on:
+      - image-build-amd
+      source_file_dependencies:
+      - .buildkite/test_areas/lm_eval.yaml
+      - tests/evals/gsm8k/
+      - tests/utils.py
+      - tests/conftest.py
+      - vllm/model_executor/layers/quantization/turboquant/
+      - vllm/v1/attention/backends/turboquant_attn.py
+      - vllm/v1/attention/backends/triton_attn.py
+      - vllm/v1/attention/ops/triton_unified_attention.py
+      - vllm/v1/attention/selector.py
+      - vllm/v1/attention/ops/triton_turboquant_decode.py
+      - vllm/v1/attention/ops/triton_turboquant_store.py
+      - vllm/v1/attention/ops/flydsl_turboquant_decode.py
+      - vllm/v1/attention/ops/flydsl_kernels/
+      - vllm/v1/attention/ops/turboquant_soa/
+      - vllm/platforms/rocm.py
+      commands:
+      - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=evals/gsm8k/configs/models-turboquant-k8v4-rocm.txt
 
 - label: ":nvidia: (H200 MIG 18GB) LM Eval TurboQuant t4nc"
   key: lm-eval-turboquant-t4nc
@@ -658,6 +686,34 @@ steps:
   - vllm/v1/attention/ops/triton_turboquant_store.py
   commands:
   - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=evals/gsm8k/configs/models-turboquant-t4nc.txt
+  mirror:
+    amd:
+      label: ":amd: (MI355) LM Eval TurboQuant t4nc"
+      dind: false
+      device: mi355_1
+      num_devices: 1
+      timeout_in_minutes: 30
+      working_dir: /vllm-workspace/tests
+      depends_on:
+      - image-build-amd
+      source_file_dependencies:
+      - .buildkite/test_areas/lm_eval.yaml
+      - tests/evals/gsm8k/
+      - tests/utils.py
+      - tests/conftest.py
+      - vllm/model_executor/layers/quantization/turboquant/
+      - vllm/v1/attention/backends/turboquant_attn.py
+      - vllm/v1/attention/backends/triton_attn.py
+      - vllm/v1/attention/ops/triton_unified_attention.py
+      - vllm/v1/attention/selector.py
+      - vllm/v1/attention/ops/triton_turboquant_decode.py
+      - vllm/v1/attention/ops/triton_turboquant_store.py
+      - vllm/v1/attention/ops/flydsl_turboquant_decode.py
+      - vllm/v1/attention/ops/flydsl_kernels/
+      - vllm/v1/attention/ops/turboquant_soa/
+      - vllm/platforms/rocm.py
+      commands:
+      - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=evals/gsm8k/configs/models-turboquant-t4nc-rocm.txt
 
 - label: ":nvidia: (H200 MIG 18GB) LM Eval TurboQuant k3v4nc"
   key: lm-eval-turboquant-k3v4nc
@@ -670,6 +726,34 @@ steps:
   - vllm/v1/attention/ops/triton_turboquant_store.py
   commands:
   - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=evals/gsm8k/configs/models-turboquant-k3v4nc.txt
+  mirror:
+    amd:
+      label: ":amd: (MI355) LM Eval TurboQuant k3v4nc"
+      dind: false
+      device: mi355_1
+      num_devices: 1
+      timeout_in_minutes: 30
+      working_dir: /vllm-workspace/tests
+      depends_on:
+      - image-build-amd
+      source_file_dependencies:
+      - .buildkite/test_areas/lm_eval.yaml
+      - tests/evals/gsm8k/
+      - tests/utils.py
+      - tests/conftest.py
+      - vllm/model_executor/layers/quantization/turboquant/
+      - vllm/v1/attention/backends/turboquant_attn.py
+      - vllm/v1/attention/backends/triton_attn.py
+      - vllm/v1/attention/ops/triton_unified_attention.py
+      - vllm/v1/attention/selector.py
+      - vllm/v1/attention/ops/triton_turboquant_decode.py
+      - vllm/v1/attention/ops/triton_turboquant_store.py
+      - vllm/v1/attention/ops/flydsl_turboquant_decode.py
+      - vllm/v1/attention/ops/flydsl_kernels/
+      - vllm/v1/attention/ops/turboquant_soa/
+      - vllm/platforms/rocm.py
+      commands:
+      - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=evals/gsm8k/configs/models-turboquant-k3v4nc-rocm.txt
 
 - label: ":nvidia: (H200 MIG 18GB) LM Eval TurboQuant t3nc"
   key: lm-eval-turboquant-t3nc
diff --git a/tests/evals/gsm8k/configs/Qwen3-4B-TQ-k3v4nc-ROCm.yaml b/tests/evals/gsm8k/configs/Qwen3-4B-TQ-k3v4nc-ROCm.yaml
new file mode 100644
index 0000000000..879ad57c20
--- /dev/null
+++ b/tests/evals/gsm8k/configs/Qwen3-4B-TQ-k3v4nc-ROCm.yaml
@@ -0,0 +1,6 @@
+model_name: "Qwen/Qwen3-4B"
+accuracy_threshold: 0.78
+num_questions: 1319
+num_fewshot: 5
+# Triton serves the unquantized boundary layers with TurboQuant-compatible layout.
+server_args: "--kv-cache-dtype turboquant_k3v4_nc --max-model-len 4096 --attention-backend TRITON_ATTN"
diff --git a/tests/evals/gsm8k/configs/Qwen3-4B-TQ-k8v4-ROCm.yaml b/tests/evals/gsm8k/configs/Qwen3-4B-TQ-k8v4-ROCm.yaml
new file mode 100644
index 0000000000..aa66561223
--- /dev/null
+++ b/tests/evals/gsm8k/configs/Qwen3-4B-TQ-k8v4-ROCm.yaml
@@ -0,0 +1,6 @@
+model_name: "Qwen/Qwen3-4B"
+accuracy_threshold: 0.80
+num_questions: 1319
+num_fewshot: 5
+# Triton serves the unquantized boundary layers with TurboQuant-compatible layout.
+server_args: "--kv-cache-dtype turboquant_k8v4 --max-model-len 4096 --attention-backend TRITON_ATTN"
diff --git a/tests/evals/gsm8k/configs/Qwen3-4B-TQ-t4nc-ROCm.yaml b/tests/evals/gsm8k/configs/Qwen3-4B-TQ-t4nc-ROCm.yaml
new file mode 100644
index 0000000000..1348661b47
--- /dev/null
+++ b/tests/evals/gsm8k/configs/Qwen3-4B-TQ-t4nc-ROCm.yaml
@@ -0,0 +1,6 @@
+model_name: "Qwen/Qwen3-4B"
+accuracy_threshold: 0.80
+num_questions: 1319
+num_fewshot: 5
+# Triton serves the unquantized boundary layers with TurboQuant-compatible layout.
+server_args: "--kv-cache-dtype turboquant_4bit_nc --max-model-len 4096 --attention-backend TRITON_ATTN"
diff --git a/tests/evals/gsm8k/configs/models-turboquant-k3v4nc-rocm.txt b/tests/evals/gsm8k/configs/models-turboquant-k3v4nc-rocm.txt
new file mode 100644
index 0000000000..9a1e252a4b
--- /dev/null
+++ b/tests/evals/gsm8k/configs/models-turboquant-k3v4nc-rocm.txt
@@ -0,0 +1 @@
+Qwen3-4B-TQ-k3v4nc-ROCm.yaml
diff --git a/tests/evals/gsm8k/configs/models-turboquant-k8v4-rocm.txt b/tests/evals/gsm8k/configs/models-turboquant-k8v4-rocm.txt
new file mode 100644
index 0000000000..458fe69f00
--- /dev/null
+++ b/tests/evals/gsm8k/configs/models-turboquant-k8v4-rocm.txt
@@ -0,0 +1 @@
+Qwen3-4B-TQ-k8v4-ROCm.yaml
diff --git a/tests/evals/gsm8k/configs/models-turboquant-t4nc-rocm.txt b/tests/evals/gsm8k/configs/models-turboquant-t4nc-rocm.txt
new file mode 100644
index 0000000000..cd599778b6
--- /dev/null
+++ b/tests/evals/gsm8k/configs/models-turboquant-t4nc-rocm.txt
@@ -0,0 +1 @@
+Qwen3-4B-TQ-t4nc-ROCm.yaml