diff --git a/src/generators.cpp b/src/generators.cpp
index cbb00b46c..5744f1653 100644
--- a/src/generators.cpp
+++ b/src/generators.cpp
@@ -28,6 +28,18 @@ OrtEnv& GetOrtEnv() {
   return *GetOrtGlobals()->env_;
 }
 
+std::string to_string(DeviceType device_type) {
+  switch (device_type) {
+    case DeviceType::CPU:
+      return "CPU";
+    case DeviceType::CUDA:
+      return "CUDA";
+    case DeviceType::DML:
+      return "DirectML";
+  }
+  throw std::runtime_error("Unknown device type");
+}
+
 GeneratorParams::GeneratorParams(const Model& model)
     : search{model.config_->search},
       pad_token_id{model.config_->model.pad_token_id},
diff --git a/src/generators.h b/src/generators.h
index bbf024004..7a8f08951 100644
--- a/src/generators.h
+++ b/src/generators.h
@@ -53,6 +53,8 @@ enum struct DeviceType {
   DML,
 };
 
+std::string to_string(DeviceType device_type);
+
 struct GeneratorParams : std::enable_shared_from_this<GeneratorParams> {
   GeneratorParams() = default;  // This constructor is only used if doing a custom model handler vs built-in
   GeneratorParams(const Model& model);
diff --git a/src/python/python.cpp b/src/python/python.cpp
index e65fb3758..02903632e 100644
--- a/src/python/python.cpp
+++ b/src/python/python.cpp
@@ -418,7 +418,7 @@ PYBIND11_MODULE(onnxruntime_genai, m) {
         return CreateModel(GetOrtEnv(), config_path.c_str());
       }))
       .def("generate", [](Model& model, PyGeneratorParams& params) { params.Prepare(); return Generate(model, params); })
-      .def_property_readonly("device_type", [](const Model& s) { return s.device_type_; })
+      .def_property_readonly("device_type", [](const Model& s) { return to_string(s.device_type_); })
       .def("create_multimodal_processor", [](const Model& model) { return model.CreateMultiModalProcessor(); });
 
   pybind11::class_<PyGenerator>(m, "Generator")
diff --git a/test/python/test_onnxruntime_genai_api.py b/test/python/test_onnxruntime_genai_api.py
index a92055434..e3e72c2c9 100644
--- a/test/python/test_onnxruntime_genai_api.py
+++ b/test/python/test_onnxruntime_genai_api.py
@@ -149,3 +149,23 @@ def test_logging():
     og.set_log_options(enabled=True, model_input_values=True, model_output_shapes=True)
     og.set_log_options(model_input_values=False, model_output_shapes=False)
     og.set_log_options(enabled=False)
+
+
+@pytest.mark.parametrize(
+    "relative_model_path",
+    (
+        [
+            (Path("hf-internal-testing") / "tiny-random-gpt2-fp32", "CPU"),
+            (Path("hf-internal-testing") / "tiny-random-gpt2-fp32-cuda", "CUDA"),
+            (Path("hf-internal-testing") / "tiny-random-gpt2-fp16-cuda", "CUDA"),
+        ]
+        if og.is_cuda_available()
+        else [(Path("hf-internal-testing") / "tiny-random-gpt2-fp32", "CPU")]
+    ),
+)
+def test_model_device_type(test_data_path, relative_model_path):
+    model_path = os.fspath(Path(test_data_path) / relative_model_path[0])
+
+    model = og.Model(model_path)
+
+    assert model.device_type == relative_model_path[1]