diff --git a/tests/integration/defs/accuracy/references/gsm8k.yaml b/tests/integration/defs/accuracy/references/gsm8k.yaml
index e2bca37c511..f8b03c1fc83 100644
--- a/tests/integration/defs/accuracy/references/gsm8k.yaml
+++ b/tests/integration/defs/accuracy/references/gsm8k.yaml
@@ -138,6 +138,11 @@ Qwen3/Qwen3-235B-A22B:
     accuracy: 85.78
 Qwen3/Qwen3-Next-80B-A3B-Thinking:
   - accuracy: 81.577
+Qwen3/Qwen3-Next-80B-A3B-Instruct:
+  - accuracy: 92.72
+  - quant_algo: NVFP4
+    kv_cache_quant_algo: FP8
+    accuracy: 90.86
 moonshotai/Kimi-K2-Instruct:
   - quant_algo: FP8_BLOCK_SCALES
     accuracy: 94.84
diff --git a/tests/integration/defs/accuracy/references/mmlu.yaml b/tests/integration/defs/accuracy/references/mmlu.yaml
index cec56c28c47..eef6b9ffa2d 100644
--- a/tests/integration/defs/accuracy/references/mmlu.yaml
+++ b/tests/integration/defs/accuracy/references/mmlu.yaml
@@ -242,6 +242,11 @@ Qwen3/Qwen3-235B-A22B:
     accuracy: 86
 Qwen3/Qwen3-Next-80B-A3B-Thinking:
   - accuracy: 86
+Qwen3/Qwen3-Next-80B-A3B-Instruct:
+  - accuracy: 86.03
+  - quant_algo: NVFP4
+    kv_cache_quant_algo: FP8
+    accuracy: 85.08
 moonshotai/Kimi-K2-Instruct:
   - quant_algo: FP8_BLOCK_SCALES
     accuracy: 87.65
diff --git a/tests/integration/defs/accuracy/test_llm_api_pytorch.py b/tests/integration/defs/accuracy/test_llm_api_pytorch.py
index af38a021c74..661107c282a 100644
--- a/tests/integration/defs/accuracy/test_llm_api_pytorch.py
+++ b/tests/integration/defs/accuracy/test_llm_api_pytorch.py
@@ -4136,6 +4136,86 @@ def test_auto_dtype(self, tp_size, pp_size, ep_size):
             task.evaluate(llm)
 
 
+@skip_pre_hopper
+@pytest.mark.skip_less_device_memory(80000)
+class TestQwen3NextInstruct(LlmapiAccuracyTestHarness):
+    MODEL_PATH = f"{llm_models_root()}/Qwen3-Next"
+    MODEL_NAME = "Qwen3/Qwen3-Next-80B-A3B-Instruct"
+
+    # Default setting of `256` is too small
+    GSM8K_MAX_OUTPUT_LEN = 512
+
+    @pytest.mark.skip_less_device(4)
+    @pytest.mark.parametrize(
+        "tp_size,pp_size,ep_size,cuda_graph,overlap_scheduler",
+        [
+            (4, 1, 4, True, True),
+        ],
+        ids=[
+            "tp4ep4_cudagraph_overlap",
+        ],
+    )
+    def test_bf16_4gpu(self, tp_size, pp_size, ep_size, cuda_graph,
+                       overlap_scheduler, mocker):
+        model_path = f"{self.MODEL_PATH}/Qwen3-Next-80B-A3B-Instruct"
+        kv_cache_config = KvCacheConfig(free_gpu_memory_fraction=0.6,
+                                        enable_block_reuse=False)
+        pytorch_config = dict(disable_overlap_scheduler=not overlap_scheduler,
+                              cuda_graph_config=CudaGraphConfig(
+                                  max_batch_size=512) if cuda_graph else None)
+
+        with LLM(
+                model_path,
+                tensor_parallel_size=tp_size,
+                max_num_tokens=16384,
+                pipeline_parallel_size=pp_size,
+                moe_expert_parallel_size=ep_size,
+                kv_cache_config=kv_cache_config,
+                **pytorch_config,
+        ) as llm:
+            task = MMLU(self.MODEL_NAME)
+            task.evaluate(llm)
+            mocker.patch.object(GSM8K, "MAX_OUTPUT_LEN",
+                                self.GSM8K_MAX_OUTPUT_LEN)
+            task = GSM8K(self.MODEL_NAME)
+            task.evaluate(llm)
+
+    @skip_pre_blackwell
+    @pytest.mark.skip_less_device(4)
+    @pytest.mark.parametrize("moe_backend", ["CUTLASS", "TRTLLM"],
+                             ids=["cutlass", "trtllm"])
+    @pytest.mark.parametrize(
+        "tp_size,pp_size,ep_size,cuda_graph,overlap_scheduler",
+        [(1, 1, 1, True, True), (4, 1, 1, True, True), (4, 1, 4, True, True),
+         (4, 1, 4, False, False)],
+        ids=["tp1", "tp4ep1", "tp4ep4", "no_cuda_graph_overlap"])
+    def test_nvfp4(self, moe_backend, tp_size, pp_size, ep_size, cuda_graph,
+                   overlap_scheduler, mocker):
+        model_path = f"{self.MODEL_PATH}/qwen3-next-80b-instruct-nvfp4-ptq-fp8kv"
+
+        kv_cache_config = KvCacheConfig(free_gpu_memory_fraction=0.6,
+                                        enable_block_reuse=False)
+        pytorch_config = dict(disable_overlap_scheduler=not overlap_scheduler,
+                              cuda_graph_config=CudaGraphConfig(
+                                  max_batch_size=512) if cuda_graph else None)
+        moe_config = MoeConfig(backend=moe_backend)
+
+        with LLM(model_path,
+                 tensor_parallel_size=tp_size,
+                 max_num_tokens=16384,
+                 pipeline_parallel_size=pp_size,
+                 moe_expert_parallel_size=ep_size,
+                 kv_cache_config=kv_cache_config,
+                 **pytorch_config,
+                 moe_config=moe_config) as llm:
+            task = MMLU(self.MODEL_NAME)
+            task.evaluate(llm)
+            mocker.patch.object(GSM8K, "MAX_OUTPUT_LEN",
+                                self.GSM8K_MAX_OUTPUT_LEN)
+            task = GSM8K(self.MODEL_NAME)
+            task.evaluate(llm)
+
+
 class TestSeedOss_36B(LlmapiAccuracyTestHarness):
     MODEL_NAME = "ByteDance-Seed/Seed-OSS-36B-Instruct"
     MODEL_PATH = f"{llm_models_root()}/Seed-OSS/Seed-OSS-36B-Instruct"
diff --git a/tests/integration/test_lists/test-db/l0_b200.yml b/tests/integration/test_lists/test-db/l0_b200.yml
index e51321d5b01..79527ee6f31 100644
--- a/tests/integration/test_lists/test-db/l0_b200.yml
+++ b/tests/integration/test_lists/test-db/l0_b200.yml
@@ -54,6 +54,11 @@ l0_b200:
   - accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B::test_w4a8_mxfp4[mxfp8-latency-TRTLLM]
   - accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B::test_w4a8_mxfp4[mxfp8-latency-CUTLASS]
   - accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B::test_w4a16_mxfp4[latency-TRTLLM]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp1-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp4ep1-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp4ep4-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[no_cuda_graph_overlap-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp4ep4-trtllm]
   - disaggregated/test_workers.py::test_workers_kv_cache_aware_router_eviction[TinyLlama-1.1B-Chat-v1.0] # nvbugs 5300551
   - test_e2e.py::test_ptp_quickstart_advanced[Llama3.1-8B-NVFP4-nvfp4-quantized/Meta-Llama-3.1-8B]
   - test_e2e.py::test_ptp_quickstart_advanced[Llama3.1-8B-FP8-llama-3.1-model/Llama-3.1-8B-Instruct-FP8]
diff --git a/tests/integration/test_lists/test-db/l0_dgx_b200.yml b/tests/integration/test_lists/test-db/l0_dgx_b200.yml
index c44ce8bbfc1..11f30e681a1 100644
--- a/tests/integration/test_lists/test-db/l0_dgx_b200.yml
+++ b/tests/integration/test_lists/test-db/l0_dgx_b200.yml
@@ -42,6 +42,12 @@ l0_dgx_b200:
   - accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B::test_nvfp4[dep4_latency_moe_cutlass-torch_compile=False]
   - accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B::test_nvfp4[dep4_latency_moe_cutlass-torch_compile=True] ISOLATION
   - accuracy/test_llm_api_pytorch.py::TestQwen3NextThinking::test_auto_dtype[tp4ep4]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_bf16_4gpu[tp4ep4_cudagraph_overlap]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp1-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp4ep1-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp4ep4-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[no_cuda_graph_overlap-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp4ep4-trtllm]
   - disaggregated/test_disaggregated.py::test_disaggregated_deepseek_v3_lite_fp8_ucx[DeepSeek-V3-Lite-fp8]
   - accuracy/test_llm_api_pytorch.py::TestGPTOSS::test_w4_4gpus[tp4-trtllm-auto]
   - accuracy/test_llm_api_pytorch.py::TestGPTOSS::test_w4_4gpus[ep4-cutlass-auto]
diff --git a/tests/integration/test_lists/test-db/l0_dgx_b300.yml b/tests/integration/test_lists/test-db/l0_dgx_b300.yml
index 749f032fedf..ecb0ec976f0 100644
--- a/tests/integration/test_lists/test-db/l0_dgx_b300.yml
+++ b/tests/integration/test_lists/test-db/l0_dgx_b300.yml
@@ -58,6 +58,11 @@ l0_dgx_b300:
   - accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B::test_nvfp4[dep4_latency_moe_trtllm-torch_compile=False]
   - accuracy/test_llm_api_pytorch.py::TestQwen3_30B_A3B::test_nvfp4[dep4_latency_moe_cutlass-torch_compile=True]
   - accuracy/test_disaggregated_serving.py::TestQwen3_30B_A3B::test_mixed_ctx_gen_model[ctxpp2gentp2]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp1-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp4ep1-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp4ep4-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[no_cuda_graph_overlap-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp4ep4-trtllm]
   - disaggregated/test_disaggregated.py::test_disaggregated_deepseek_v3_lite_fp8_ucx[DeepSeek-V3-Lite-fp8]
   - accuracy/test_llm_api_pytorch.py::TestGPTOSS::test_w4_4gpus[tp4-cutlass-fp8]
   - accuracy/test_llm_api_pytorch.py::TestGPTOSS::test_w4_4gpus[tp4-cutlass-auto]
diff --git a/tests/integration/test_lists/test-db/l0_dgx_h100.yml b/tests/integration/test_lists/test-db/l0_dgx_h100.yml
index 9a875ff99d8..7e50c6ebf88 100644
--- a/tests/integration/test_lists/test-db/l0_dgx_h100.yml
+++ b/tests/integration/test_lists/test-db/l0_dgx_h100.yml
@@ -93,6 +93,7 @@ l0_dgx_h100:
   - accuracy/test_disaggregated_serving.py::TestLlama3_1_8BInstruct::test_ctx_pp_gen_tp_asymmetric[MMLU-gen_tp=2-ctx_pp=2]
   - accuracy/test_disaggregated_serving.py::TestLlama3_1_8BInstruct::test_multi_instance[GSM8K]
   - accuracy/test_disaggregated_serving.py::TestLlama3_1_8BInstruct::test_multi_instance[MMLU]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_bf16_4gpu[tp4ep4_cudagraph_overlap]
   - disaggregated/test_auto_scaling.py::test_service_discovery[etcd-round_robin]
   - disaggregated/test_auto_scaling.py::test_service_discovery[etcd-load_balancing]
   - disaggregated/test_auto_scaling.py::test_worker_restart[etcd-round_robin]
diff --git a/tests/integration/test_lists/test-db/l0_dgx_h200.yml b/tests/integration/test_lists/test-db/l0_dgx_h200.yml
index efacea5814d..d0e45202d9f 100644
--- a/tests/integration/test_lists/test-db/l0_dgx_h200.yml
+++ b/tests/integration/test_lists/test-db/l0_dgx_h200.yml
@@ -33,6 +33,7 @@ l0_dgx_h200:
   - accuracy/test_disaggregated_serving.py::TestLlama3_1_8BInstruct::test_ctx_pp_gen_tp_asymmetric[MMLU-gen_tp=2-ctx_pp=4]
   - accuracy/test_disaggregated_serving.py::TestGPTOSS::test_auto_dtype[False]
   - accuracy/test_disaggregated_serving.py::TestGPTOSS::test_auto_dtype[True]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_bf16_4gpu[tp4ep4_cudagraph_overlap]
   - disaggregated/test_disaggregated.py::test_disaggregated_ctxtp2pp2_gentp2pp2[TinyLlama-1.1B-Chat-v1.0]
   - disaggregated/test_disaggregated.py::test_disaggregated_ctxpp4_genpp4[TinyLlama-1.1B-Chat-v1.0]
   - unittest/llmapi/test_llm_pytorch.py::test_nemotron_nas_lora
diff --git a/tests/integration/test_lists/test-db/l0_gb200_multi_gpus.yml b/tests/integration/test_lists/test-db/l0_gb200_multi_gpus.yml
index 3898b101ea3..9c8d442401a 100644
--- a/tests/integration/test_lists/test-db/l0_gb200_multi_gpus.yml
+++ b/tests/integration/test_lists/test-db/l0_gb200_multi_gpus.yml
@@ -43,6 +43,11 @@ l0_gb200_multi_gpus:
   - accuracy/test_llm_api_pytorch.py::TestGPTOSS::test_w4_4gpus[tp4-trtllm-auto]
   - accuracy/test_llm_api_pytorch.py::TestGPTOSS::test_w4_4gpus[dp4-cutlass-auto]
   - accuracy/test_llm_api_pytorch.py::TestGPTOSS::test_w4_4gpus_online_eplb[fp8]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp1-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp4ep1-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp4ep4-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[no_cuda_graph_overlap-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp4ep4-trtllm]
 - condition:
     ranges:
       system_gpu_count:
diff --git a/tests/integration/test_lists/test-db/l0_rtx_pro_6000.yml b/tests/integration/test_lists/test-db/l0_rtx_pro_6000.yml
index 786b03e0e38..ececeab4ae8 100644
--- a/tests/integration/test_lists/test-db/l0_rtx_pro_6000.yml
+++ b/tests/integration/test_lists/test-db/l0_rtx_pro_6000.yml
@@ -43,6 +43,8 @@ l0_rtx_pro_6000:
   - accuracy/test_llm_api_pytorch.py::TestGPTOSS::test_w4_1gpu[True-True-cutlass-auto]
   - accuracy/test_llm_api_pytorch.py::TestPhi4MM::test_fp4
   - accuracy/test_llm_api_pytorch.py::TestPhi4MM::test_fp8
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp1-cutlass]
+  - accuracy/test_llm_api_pytorch.py::TestQwen3NextInstruct::test_nvfp4[tp4ep4-trtllm]
 
 - condition:
     ranges: