ai-agents-azure-openai-streami-streaming-with-azurechatopenai-using-responses-api-59b8a2d1

THEN Use the invoke method instead of stream to get the response including reasoning summary. For example, call `az_llm.invoke(messages)` instead of iterating over `chain.stream()`. Avoid streaming when using the Responses API with reasoning until the underlying issue is resolved.

az_llm = AzureChatOpenAI(azure_endpoint=..., model_name='o4-mini', use_responses_api=True, model_kwargs={'reasoning': {'effort': 'medium'}})
messages = [HumanMessage(content='Tell me about neural text degeneration')]
response = az_llm.invoke(messages)
print(response.content)