diff --git a/google/genai/_live_converters.py b/google/genai/_live_converters.py
index 5512f3876..aeb66e919 100644
--- a/google/genai/_live_converters.py
+++ b/google/genai/_live_converters.py
@@ -1674,6 +1674,11 @@ def _Part_to_mldev(
         getv(from_object, ['audio_transcription']),
     )
 
+  if getv(from_object, ['media_processing']) is not None:
+    setv(
+        to_object, ['mediaProcessing'], getv(from_object, ['media_processing'])
+    )
+
   return to_object
 
 
@@ -1762,6 +1767,11 @@ def _Part_to_vertex(
         getv(from_object, ['audio_transcription']),
     )
 
+  if getv(from_object, ['media_processing']) is not None:
+    setv(
+        to_object, ['mediaProcessing'], getv(from_object, ['media_processing'])
+    )
+
   return to_object
 
 
diff --git a/google/genai/_tokens_converters.py b/google/genai/_tokens_converters.py
index 8ece00fb5..a331e0c12 100644
--- a/google/genai/_tokens_converters.py
+++ b/google/genai/_tokens_converters.py
@@ -577,6 +577,11 @@ def _Part_to_mldev(
         getv(from_object, ['audio_transcription']),
     )
 
+  if getv(from_object, ['media_processing']) is not None:
+    setv(
+        to_object, ['mediaProcessing'], getv(from_object, ['media_processing'])
+    )
+
   return to_object
 
 
diff --git a/google/genai/batches.py b/google/genai/batches.py
index a01715d59..1e8a424ff 100644
--- a/google/genai/batches.py
+++ b/google/genai/batches.py
@@ -1626,6 +1626,11 @@ def _Part_to_mldev(
         getv(from_object, ['audio_transcription']),
     )
 
+  if getv(from_object, ['media_processing']) is not None:
+    setv(
+        to_object, ['mediaProcessing'], getv(from_object, ['media_processing'])
+    )
+
   return to_object
 
 
diff --git a/google/genai/caches.py b/google/genai/caches.py
index 90c77e897..91f02bc6f 100644
--- a/google/genai/caches.py
+++ b/google/genai/caches.py
@@ -802,6 +802,11 @@ def _Part_to_mldev(
         getv(from_object, ['audio_transcription']),
     )
 
+  if getv(from_object, ['media_processing']) is not None:
+    setv(
+        to_object, ['mediaProcessing'], getv(from_object, ['media_processing'])
+    )
+
   return to_object
 
 
@@ -890,6 +895,11 @@ def _Part_to_vertex(
         getv(from_object, ['audio_transcription']),
     )
 
+  if getv(from_object, ['media_processing']) is not None:
+    setv(
+        to_object, ['mediaProcessing'], getv(from_object, ['media_processing'])
+    )
+
   return to_object
 
 
diff --git a/google/genai/models.py b/google/genai/models.py
index 1d8dbd8d6..0e4fc3b7e 100644
--- a/google/genai/models.py
+++ b/google/genai/models.py
@@ -3708,6 +3708,11 @@ def _Part_to_mldev(
         getv(from_object, ['audio_transcription']),
     )
 
+  if getv(from_object, ['media_processing']) is not None:
+    setv(
+        to_object, ['mediaProcessing'], getv(from_object, ['media_processing'])
+    )
+
   return to_object
 
 
@@ -3797,6 +3802,11 @@ def _Part_to_vertex(
         getv(from_object, ['audio_transcription']),
     )
 
+  if getv(from_object, ['media_processing']) is not None:
+    setv(
+        to_object, ['mediaProcessing'], getv(from_object, ['media_processing'])
+    )
+
   return to_object
 
 
diff --git a/google/genai/tunings.py b/google/genai/tunings.py
index 13d9283e6..236022ae0 100644
--- a/google/genai/tunings.py
+++ b/google/genai/tunings.py
@@ -1588,6 +1588,11 @@ def _Part_to_vertex(
         getv(from_object, ['audio_transcription']),
     )
 
+  if getv(from_object, ['media_processing']) is not None:
+    setv(
+        to_object, ['mediaProcessing'], getv(from_object, ['media_processing'])
+    )
+
   return to_object
 
 
diff --git a/google/genai/types.py b/google/genai/types.py
index 78266988b..c4cb44af0 100644
--- a/google/genai/types.py
+++ b/google/genai/types.py
@@ -1013,6 +1013,17 @@ class ComputationBasedMetricType(_common.CaseInSensitiveEnum):
   """ROUGE metric."""
 
 
+class MediaProcessing(_common.CaseInSensitiveEnum):
+  """How the model processes input media for understanding."""
+
+  MEDIA_PROCESSING_UNSPECIFIED = 'MEDIA_PROCESSING_UNSPECIFIED'
+  """Default. Uses model-specific processing (3.5 Pro+ -> `AGENTIC`, older models -> `STATIC`)."""
+  STATIC = 'STATIC'
+  """Fixed-rate frame extraction. All frames placed in context."""
+  AGENTIC = 'AGENTIC'
+  """Model-driven dynamic navigation. Recommended for most use cases."""
+
+
 class PartMediaResolutionLevel(_common.CaseInSensitiveEnum):
   """The tokenization quality used for given media."""
 
@@ -2291,6 +2302,10 @@ class Part(_common.BaseModel):
       default=None,
       description="""Output only. The transcription of the audio part.""",
   )
+  media_processing: Optional[MediaProcessing] = Field(
+      default=None,
+      description="""How the model processes this part's media for understanding.""",
+  )
 
   def __init__(
       self,
@@ -2525,6 +2540,9 @@ class PartDict(TypedDict, total=False):
   audio_transcription: Optional[TranscriptionDict]
   """Output only. The transcription of the audio part."""
 
+  media_processing: Optional[MediaProcessing]
+  """How the model processes this part's media for understanding."""
+
 
 PartOrDict = Union[Part, PartDict]