https://blog.metakode.co/articles/openai-launches-flex-processing-for-cheaper--slower-ai-tasks