model = "outsource/openai-chat/gpt-4-turbo"
prompt_template_id = "codegen"
max_tokens = 2048