OpenAI Just Changed the Economics of Workflow Automation!

What if you can automate a task for $0.01? Actually even less, we are just rounding up!

How did OpenAI do this? Simple. Not every task needs a professor. Some tasks can be completed by a high-school level student. With the introduction of Luna, Terra and Sol models, OpenAI has introduced AI assistants of varying skills at different price points. Now you can pick the right model for the task to be automated.

In general, I am not a big fan of giving users more knobs to turn. And OpenAI took away some knobs in a previous ChatGPT release by auto-selecting the right level of intelligence based on the query. But for enterprise workflow automation, enabling model choice is a fantastic idea. It enables enterprise customers to choose the right AI model for the task at hand.

We have been so focused on the latest, greatest model that we were a little surprised that the smaller (?), definitely cheaper Luna model performed the best for one of our simple automation tasks. See results below. Luna is cheaper and faster than Sol, Opus 5.7 and Kimi 3!

At $0.01 for this task to be automated, ROI is no longer a question! GPT 5.6 Luna has completely changed the economics of workflow automation.

As we explore more customer workflows, we will likely need more expensive models in many cases. We will have more to share in the coming weeks and months.