infrastructure-azure-openai-respons-when-using-the-openai-official-client-library-to-c-8bcb0f20

THEN Use the LiteLLM client instead, or upgrade to LiteLLM version 1.75.9 or later which includes the fix from PR #13526. Alternatively, use the /chat/completions endpoint which works correctly with the official client.

import openai
client = openai.OpenAI(base_url='<litellm_url>', api_key='<api_key>')
# This fails for Azure OpenAI models:
# client.responses.create(model='gpt-4.1-mini', input=[{'role':'user','content':'Say OK!'}])
# Use instead:
client.chat.completions.create(model='gpt-4.1-mini', messages=[{'role':'user','content':'Say OK!'}])