4v-think

Author: jinhaiyang

tests/integration_tests/test_vlm_thinking.py

@@ -0,0 +1,97 @@
+import os.path
+
+from zhipuai import ZhipuAI
+import zhipuai
+import time
+
+import logging
+import logging.config
+
+
+def test_completions_vlm_thinking(logging_conf):
+    logging.config.dictConfig(logging_conf)  # type: ignore
+    client = ZhipuAI()  # 填写您自己的APIKey
+    try:
+        # 生成request_id
+        request_id = time.time()
+        print(f"request_id:{request_id}")
+        response = client.chat.completions.create(
+            request_id=request_id,
+            model="glm-4.1v-thinking-flash",  # 填写需要调用的模型名称
+            messages=[
+                {
+                    "role": "user",
+                    "content": [
+                        {
+                            "type": "text",
+                            "text": "图里有什么"
+                        },
+
+                        {
+                            "type": "image_url",
+                            "image_url": {
+                                "url": "https://img1.baidu.com/it/u=1369931113,3388870256&fm=253&app=138&size=w931&n=0&f=JPEG&fmt=auto?sec=1703696400&t=f3028c7a1dca43a080aeb8239f09cc2f"
+                            }
+                        }
+                    ]
+                }
+            ],
+            temperature=0.5,
+            max_tokens = 1024,
+            user_id="12345678"
+        )
+        print(response)
+
+
+    except zhipuai.core._errors.APIRequestFailedError as err:
+        print(err)
+    except zhipuai.core._errors.APIInternalError as err:
+        print(err)
+    except zhipuai.core._errors.APIStatusError as err:
+        print(err)
+
+
+
+def test_completions_vlm_thinking_stream(logging_conf):
+    logging.config.dictConfig(logging_conf)  # type: ignore
+    client = ZhipuAI()  # 填写您自己的APIKey
+    try:
+        # 生成request_id
+        request_id = time.time()
+        print(f"request_id:{request_id}")
+        response = client.chat.completions.create(
+            request_id=request_id,
+            model="glm-4.1v-thinking-flash",  # 填写需要调用的模型名称
+            messages=[
+                {
+                    "role": "user",
+                    "content": [
+                        {
+                            "type": "text",
+                            "text": "图里有什么"
+                        },
+
+                        {
+                            "type": "image_url",
+                            "image_url": {
+                                "url": "https://img1.baidu.com/it/u=1369931113,3388870256&fm=253&app=138&size=w931&n=0&f=JPEG&fmt=auto?sec=1703696400&t=f3028c7a1dca43a080aeb8239f09cc2f"
+                            }
+                        }
+                    ]
+                }
+            ],
+            temperature=0.5,
+            max_tokens = 1024,
+            user_id="12345678",
+            stream=True
+        )
+        for item in response:
+            print(item)
+
+    except zhipuai.core._errors.APIRequestFailedError as err:
+        print(err)
+    except zhipuai.core._errors.APIInternalError as err:
+        print(err)
+    except zhipuai.core._errors.APIStatusError as err:
+        print(err)
+

zhipuai/types/chat/chat_completion.py

@@ -19,6 +19,7 @@ class CompletionMessageToolCall(BaseModel):
 class CompletionMessage(BaseModel):
     content: Optional[str] = None
     role: str
+    reasoning_content: Optional[str] = None
     tool_calls: Optional[List[CompletionMessageToolCall]] = None
 
 

zhipuai/types/chat/chat_completion_chunk.py

@@ -39,6 +39,7 @@ class AudioCompletionChunk(BaseModel):
 class ChoiceDelta(BaseModel):
     content: Optional[str] = None
     role: Optional[str] = None
+    reasoning_content: Optional[str] = None
     tool_calls: Optional[List[ChoiceDeltaToolCall]] = None
     audio: Optional[AudioCompletionChunk] = None