diff --git a/src/diffusers/pipelines/flux/pipeline_flux_controlnet.py b/src/diffusers/pipelines/flux/pipeline_flux_controlnet.py
index 0ce8628c0822..eee41b9af4d1 100644
--- a/src/diffusers/pipelines/flux/pipeline_flux_controlnet.py
+++ b/src/diffusers/pipelines/flux/pipeline_flux_controlnet.py
@@ -202,7 +202,7 @@ class FluxControlNetPipeline(DiffusionPipeline, FluxLoraLoaderMixin, FromSingleF
 
     model_cpu_offload_seq = "text_encoder->text_encoder_2->image_encoder->transformer->vae"
     _optional_components = ["image_encoder", "feature_extractor"]
-    _callback_tensor_inputs = ["latents", "prompt_embeds"]
+    _callback_tensor_inputs = ["latents", "prompt_embeds", "control_image"]
 
     def __init__(
         self,
@@ -1149,6 +1149,7 @@ def __call__(
 
                     latents = callback_outputs.pop("latents", latents)
                     prompt_embeds = callback_outputs.pop("prompt_embeds", prompt_embeds)
+                    control_image = callback_outputs.pop("control_image", control_image)
 
                 # call the callback, if provided
                 if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):
diff --git a/src/diffusers/pipelines/flux/pipeline_flux_controlnet_image_to_image.py b/src/diffusers/pipelines/flux/pipeline_flux_controlnet_image_to_image.py
index 37b4b2657346..6219662b496f 100644
--- a/src/diffusers/pipelines/flux/pipeline_flux_controlnet_image_to_image.py
+++ b/src/diffusers/pipelines/flux/pipeline_flux_controlnet_image_to_image.py
@@ -198,7 +198,7 @@ class FluxControlNetImg2ImgPipeline(DiffusionPipeline, FluxLoraLoaderMixin, From
 
     model_cpu_offload_seq = "text_encoder->text_encoder_2->transformer->vae"
     _optional_components = []
-    _callback_tensor_inputs = ["latents", "prompt_embeds"]
+    _callback_tensor_inputs = ["latents", "prompt_embeds", "control_image"]
 
     def __init__(
         self,
@@ -973,6 +973,7 @@ def __call__(
 
                     latents = callback_outputs.pop("latents", latents)
                     prompt_embeds = callback_outputs.pop("prompt_embeds", prompt_embeds)
+                    control_image = callback_outputs.pop("control_image", control_image)
 
                 if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):
                     progress_bar.update()
diff --git a/src/diffusers/pipelines/flux/pipeline_flux_controlnet_inpainting.py b/src/diffusers/pipelines/flux/pipeline_flux_controlnet_inpainting.py
index 480e441d15ed..4d43ccd318d5 100644
--- a/src/diffusers/pipelines/flux/pipeline_flux_controlnet_inpainting.py
+++ b/src/diffusers/pipelines/flux/pipeline_flux_controlnet_inpainting.py
@@ -200,7 +200,7 @@ class FluxControlNetInpaintPipeline(DiffusionPipeline, FluxLoraLoaderMixin, From
 
     model_cpu_offload_seq = "text_encoder->text_encoder_2->transformer->vae"
     _optional_components = []
-    _callback_tensor_inputs = ["latents", "prompt_embeds"]
+    _callback_tensor_inputs = ["latents", "prompt_embeds", "control_image", "mask", "masked_image_latents"]
 
     def __init__(
         self,
@@ -1178,6 +1178,9 @@ def __call__(
 
                     latents = callback_outputs.pop("latents", latents)
                     prompt_embeds = callback_outputs.pop("prompt_embeds", prompt_embeds)
+                    control_image = callback_outputs.pop("control_image", control_image)
+                    mask = callback_outputs.pop("mask", mask)
+                    masked_image_latents = callback_outputs.pop("masked_image_latents", masked_image_latents)
 
                 if i == len(timesteps) - 1 or ((i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0):
                     progress_bar.update()