diff --git a/backends/arm/test/ops/test_argmax.py b/backends/arm/test/ops/test_argmax.py
index 08e2172cdeb..f29137fcf47 100644
--- a/backends/arm/test/ops/test_argmax.py
+++ b/backends/arm/test/ops/test_argmax.py
@@ -5,12 +5,14 @@
 
 from typing import Tuple
 
+import pytest
 import torch
 from executorch.backends.arm.test import common
 from executorch.backends.arm.test.tester.test_pipeline import (
     OpNotSupportedPipeline,
     TosaPipelineFP,
     TosaPipelineINT,
+    VgfPipeline,
 )
 
 aten_op = "torch.ops.aten.argmax.default"
@@ -131,3 +133,49 @@ def test_argmax_tosa_INT(test_data: Tuple[input_t, int]):
     )
     pipeline.count_tosa_ops({"ARGMAX": 1})
     pipeline.run()
+
+
+@pytest.mark.xfail(
+    reason=(
+        "VGF ARGMAX output contract mismatch: MLTORCH-2372"
+        "ExecuTorch output is allocated as int64 while TOSA/VGF emits int32."
+    ),
+    strict=True,
+)
+@common.parametrize(
+    "test_data",
+    Argmax.test_data | Argmax.test_data_fp16 | Argmax.test_data_bf16,
+)
+@common.SkipIfNoModelConverter
+def test_argmax_vgf_no_quant(test_data: Tuple[input_t, int]):
+    data, dim = test_data()
+    pipeline = VgfPipeline[input_t](
+        Argmax(dim),
+        data,
+        aten_op,
+        exir_op,
+        quantize=False,
+        tosa_extensions=["bf16"],
+    )
+    pipeline.run()
+
+
+@pytest.mark.xfail(
+    reason=(
+        "VGF ARGMAX output contract mismatch: MLTORCH-2372"
+        "ExecuTorch output is allocated as int64 while TOSA/VGF emits int32."
+    ),
+    strict=True,
+)
+@common.parametrize("test_data", Argmax.test_data)
+@common.SkipIfNoModelConverter
+def test_argmax_vgf_quant(test_data: Tuple[input_t, int]):
+    data, dim = test_data()
+    pipeline = VgfPipeline[input_t](
+        Argmax(dim),
+        data,
+        aten_op,
+        exir_op,
+        quantize=True,
+    )
+    pipeline.run()
diff --git a/backends/arm/test/ops/test_pow.py b/backends/arm/test/ops/test_pow.py
index 2cde848c7ef..7bec41e81fb 100644
--- a/backends/arm/test/ops/test_pow.py
+++ b/backends/arm/test/ops/test_pow.py
@@ -57,6 +57,21 @@ class Pow_TensorTensor(torch.nn.Module):
         ),
     }
 
+    test_data_quant = {
+        "positive_base_positive_exp": lambda: (
+            torch.rand(2, 3, 4) * 1.5 + 0.25,
+            torch.rand(2, 3, 4) * 1.5 + 0.25,
+        ),
+        "broadcast_exp": lambda: (
+            torch.rand(1, 3, 4, 4) * 1.5 + 0.25,
+            torch.rand(1, 3, 1, 1) * 1.5 + 0.25,
+        ),
+        "integer_valued_exp": lambda: (
+            torch.rand(2, 4) * 1.5 + 0.25,
+            torch.full((2, 4), 2.0),
+        ),
+    }
+
     def forward(self, x: torch.Tensor | float, y: torch.Tensor | float):
         return torch.pow(x, y)
 
@@ -238,3 +253,24 @@ def test_pow_tensor_scalar_vgf_quant(test_data: Pow_TensorScalar.input_t):
         quantize=True,
     )
     pipeline.run()
+
+
+@common.parametrize("test_data", Pow_TensorTensor.test_data_quant)
+@common.SkipIfNoModelConverter
+def test_pow_tensor_tensor_vgf_quant(
+    test_data: Pow_TensorTensor.input_t,
+):
+    pipeline = VgfPipeline[Pow_TensorTensor.input_t](
+        Pow_TensorTensor(),
+        test_data(),
+        Pow_TensorTensor.aten_op,
+        Pow_TensorTensor.exir_op,
+        quantize=True,
+    )
+
+    pipeline.run_and_compare_to_initial_model(
+        frobenius_threshold=0.3,
+        cosine_threshold=0.9,
+    )
+
+    pipeline.run()
diff --git a/backends/arm/test/ops/test_sign.py b/backends/arm/test/ops/test_sign.py
index 5e5a88011e9..95ec2f30797 100644
--- a/backends/arm/test/ops/test_sign.py
+++ b/backends/arm/test/ops/test_sign.py
@@ -17,7 +17,7 @@
 )
 
 aten_op = "torch.ops.aten.sign.default"
-exir_op = "executorch_exir_dialects_edge__ops_aten__sign_default"
+exir_op = "executorch_exir_dialects_edge__ops_aten_sign_default"
 
 input_t1 = Tuple[torch.Tensor]