Apply patch [skip ci]

libs/partners/openai/test_gpt5_compatibility.py

@@ -0,0 +1,27 @@
+from langchain_openai import AzureChatOpenAI
+from pydantic import SecretStr
+
+# Test the original issue scenario
+chat = AzureChatOpenAI(
+    azure_deployment="gpt-5-nano",
+    api_version="2024-12-01-preview",
+    azure_endpoint="https://example.com/",
+    api_key=SecretStr("test-key"),
+    max_completion_tokens=1500
+)
+
+print("✅ max_completion_tokens parameter accepted successfully")
+print(f"max_tokens field value: {chat.max_tokens}")
+
+# Also test max_tokens still works
+chat2 = AzureChatOpenAI(
+    azure_deployment="gpt-5-nano",
+    api_version="2024-12-01-preview",
+    azure_endpoint="https://example.com/",
+    api_key=SecretStr("test-key"),
+    max_tokens=1500
+)
+
+print("✅ max_tokens parameter still works")
+print(f"max_tokens field value: {chat2.max_tokens}")
+print("\n🎉 GPT-5 compatibility issue resolved!")