diff --git a/ollama/_types.py b/ollama/_types.py
index 8931cea..709a40e 100644
--- a/ollama/_types.py
+++ b/ollama/_types.py
@@ -551,7 +551,7 @@ class ShowResponse(SubscriptableBaseModel):
 
   details: Optional[ModelDetails] = None
 
-  modelinfo: Optional[Mapping[str, Any]] = Field(alias='model_info')
+  modelinfo: Optional[Mapping[str, Any]] = Field(default=None, alias='model_info')
 
   parameters: Optional[str] = None
 
diff --git a/tests/test_type_serialization.py b/tests/test_type_serialization.py
index f458cd2..3d9747c 100644
--- a/tests/test_type_serialization.py
+++ b/tests/test_type_serialization.py
@@ -4,7 +4,7 @@
 
 import pytest
 
-from ollama._types import CreateRequest, Image
+from ollama._types import CreateRequest, Image, ShowResponse
 
 
 def test_image_serialization_bytes():
@@ -92,3 +92,59 @@ def test_create_request_serialization_license_list():
   request = CreateRequest(model='test-model', license=['MIT', 'Apache-2.0'])
   serialized = request.model_dump()
   assert serialized['license'] == ['MIT', 'Apache-2.0']
+
+
+def test_show_response_without_model_info():
+  """
+  Test that ShowResponse can be created without model_info field.
+
+  This is a regression test for issue #607 where certain cloud models
+  (e.g., glm-4.7:cloud, qwen3-next:80b-cloud, deepseek-v3.2:cloud) return
+  responses without the model_info field, causing a ValidationError.
+  """
+  # Response data without model_info field (as returned by some cloud models)
+  response_data = {
+    'modelfile': '# Modelfile generated by "ollama show"',
+    'template': '{{ .Prompt }}',
+    'details': {
+      'parent_model': '',
+      'format': 'gguf',
+      'family': 'glm',
+      'families': ['glm'],
+      'parameter_size': '9.4B',
+      'quantization_level': 'Q4_K_M',
+    },
+    'capabilities': ['completion'],
+    'modified_at': '2025-01-01T00:00:00Z',
+  }
+
+  # This should not raise a ValidationError
+  response = ShowResponse.model_validate(response_data)
+
+  assert response.modelfile == '# Modelfile generated by "ollama show"'
+  assert response.template == '{{ .Prompt }}'
+  assert response.modelinfo is None  # model_info was not provided
+  assert response.capabilities == ['completion']
+
+
+def test_show_response_with_model_info():
+  """
+  Test that ShowResponse still works correctly when model_info is provided.
+  """
+  response_data = {
+    'modelfile': '# Modelfile',
+    'template': '{{ .Prompt }}',
+    'model_info': {
+      'general.architecture': 'llama',
+      'general.parameter_count': 7000000000,
+    },
+    'capabilities': ['completion'],
+    'modified_at': '2025-01-01T00:00:00Z',
+  }
+
+  response = ShowResponse.model_validate(response_data)
+
+  assert response.modelfile == '# Modelfile'
+  assert response.modelinfo is not None
+  assert response.modelinfo['general.architecture'] == 'llama'
+  assert response.modelinfo['general.parameter_count'] == 7000000000