diff --git a/src/diffusers/modular_pipelines/qwenimage/modular_blocks_qwenimage.py b/src/diffusers/modular_pipelines/qwenimage/modular_blocks_qwenimage.py
index 63e9f5a28372..ebe0bbbd75ba 100644
--- a/src/diffusers/modular_pipelines/qwenimage/modular_blocks_qwenimage.py
+++ b/src/diffusers/modular_pipelines/qwenimage/modular_blocks_qwenimage.py
@@ -12,9 +12,14 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+from typing import List
+
+import PIL.Image
+import torch
+
 from ...utils import logging
 from ..modular_pipeline import AutoPipelineBlocks, ConditionalPipelineBlocks, SequentialPipelineBlocks
-from ..modular_pipeline_utils import InsertableDict
+from ..modular_pipeline_utils import InsertableDict, OutputParam
 from .before_denoise import (
     QwenImageControlNetBeforeDenoiserStep,
     QwenImageCreateMaskLatentsStep,
@@ -394,6 +399,14 @@ def description(self):
             + " - for text-to-image generation, all you need to provide is prompt embeddings"
         )
 
+    @property
+    def outputs(self):
+        return [
+            OutputParam(
+                name="latents", type_hint=torch.Tensor, description="The latents generated by the denoising step"
+            ),
+        ]
+
 
 # ====================
 # 3. DECODE
@@ -467,3 +480,9 @@ def description(self):
             + "- to run the controlnet workflow, you need to provide `control_image`\n"
             + "- for text-to-image generation, all you need to provide is `prompt`"
         )
+
+    @property
+    def outputs(self):
+        return [
+            OutputParam(name="images", type_hint=List[List[PIL.Image.Image]]),
+        ]
diff --git a/src/diffusers/modular_pipelines/qwenimage/modular_blocks_qwenimage_edit.py b/src/diffusers/modular_pipelines/qwenimage/modular_blocks_qwenimage_edit.py
index 99a349994c19..2683e64080bf 100644
--- a/src/diffusers/modular_pipelines/qwenimage/modular_blocks_qwenimage_edit.py
+++ b/src/diffusers/modular_pipelines/qwenimage/modular_blocks_qwenimage_edit.py
@@ -12,11 +12,14 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from typing import Optional
+from typing import List, Optional
+
+import PIL.Image
+import torch
 
 from ...utils import logging
 from ..modular_pipeline import AutoPipelineBlocks, ConditionalPipelineBlocks, SequentialPipelineBlocks
-from ..modular_pipeline_utils import InsertableDict
+from ..modular_pipeline_utils import InsertableDict, OutputParam
 from .before_denoise import (
     QwenImageCreateMaskLatentsStep,
     QwenImageEditRoPEInputsStep,
@@ -307,6 +310,14 @@ def description(self):
             " - `QwenImageEditDecodeStep` (edit) is used when `mask` is not provided.\n"
         )
 
+    @property
+    def outputs(self):
+        return [
+            OutputParam(
+                name="latents", type_hint=torch.Tensor, description="The latents generated by the denoising step"
+            ),
+        ]
+
 
 # ====================
 # 5. AUTO BLOCKS & PRESETS
@@ -334,3 +345,9 @@ def description(self):
             "- for edit (img2img) generation, you need to provide `image`\n"
             "- for edit inpainting, you need to provide `mask_image` and `image`, optionally you can provide `padding_mask_crop`\n"
         )
+
+    @property
+    def outputs(self):
+        return [
+            OutputParam(name="images", type_hint=List[List[PIL.Image.Image]], description="The generated images"),
+        ]
diff --git a/src/diffusers/modular_pipelines/qwenimage/modular_blocks_qwenimage_edit_plus.py b/src/diffusers/modular_pipelines/qwenimage/modular_blocks_qwenimage_edit_plus.py
index 275e4288eb0a..99c5b109bf38 100644
--- a/src/diffusers/modular_pipelines/qwenimage/modular_blocks_qwenimage_edit_plus.py
+++ b/src/diffusers/modular_pipelines/qwenimage/modular_blocks_qwenimage_edit_plus.py
@@ -12,9 +12,14 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+from typing import List
+
+import PIL.Image
+import torch
+
 from ...utils import logging
 from ..modular_pipeline import SequentialPipelineBlocks
-from ..modular_pipeline_utils import InsertableDict
+from ..modular_pipeline_utils import InsertableDict, OutputParam
 from .before_denoise import (
     QwenImageEditPlusRoPEInputsStep,
     QwenImagePrepareLatentsStep,
@@ -136,6 +141,14 @@ class QwenImageEditPlusCoreDenoiseStep(SequentialPipelineBlocks):
     def description(self):
         return "Core denoising workflow for QwenImage-Edit Plus edit (img2img) task."
 
+    @property
+    def outputs(self):
+        return [
+            OutputParam(
+                name="latents", type_hint=torch.Tensor, description="The latents generated by the denoising step"
+            ),
+        ]
+
 
 # ====================
 # 4. DECODE
@@ -179,3 +192,9 @@ def description(self):
             "- Each image is resized independently based on its own aspect ratio.\n"
             "- VL encoder uses 384x384 target area, VAE encoder uses 1024x1024 target area."
         )
+
+    @property
+    def outputs(self):
+        return [
+            OutputParam(name="images", type_hint=List[List[PIL.Image.Image]], description="The generated images"),
+        ]
diff --git a/src/diffusers/modular_pipelines/qwenimage/modular_blocks_qwenimage_layered.py b/src/diffusers/modular_pipelines/qwenimage/modular_blocks_qwenimage_layered.py
index fe6f756789af..63ee36df5112 100644
--- a/src/diffusers/modular_pipelines/qwenimage/modular_blocks_qwenimage_layered.py
+++ b/src/diffusers/modular_pipelines/qwenimage/modular_blocks_qwenimage_layered.py
@@ -13,9 +13,14 @@
 # limitations under the License.
 
 
+from typing import List
+
+import PIL.Image
+import torch
+
 from ...utils import logging
 from ..modular_pipeline import SequentialPipelineBlocks
-from ..modular_pipeline_utils import InsertableDict
+from ..modular_pipeline_utils import InsertableDict, OutputParam
 from .before_denoise import (
     QwenImageLayeredPrepareLatentsStep,
     QwenImageLayeredRoPEInputsStep,
@@ -134,6 +139,14 @@ class QwenImageLayeredCoreDenoiseStep(SequentialPipelineBlocks):
     def description(self):
         return "Core denoising workflow for QwenImage-Layered img2img task."
 
+    @property
+    def outputs(self):
+        return [
+            OutputParam(
+                name="latents", type_hint=torch.Tensor, description="The latents generated by the denoising step"
+            ),
+        ]
+
 
 # ====================
 # 4. AUTO BLOCKS & PRESETS
@@ -157,3 +170,9 @@ class QwenImageLayeredAutoBlocks(SequentialPipelineBlocks):
     @property
     def description(self):
         return "Auto Modular pipeline for layered denoising tasks using QwenImage-Layered."
+
+    @property
+    def outputs(self):
+        return [
+            OutputParam(name="images", type_hint=List[List[PIL.Image.Image]], description="The generated images"),
+        ]