ChatGPT’s Big Upgrade

By: Sarah Pita

Last week, with much fanfare (at least if you were listening for it) OpenAI launched GPT-5 on ChatGPT, and it has been a bit chaotic ever since. 

When they rolled it out, all the “old” (as in, produced in the last 2+ years) models of ChatGPT were replaced by GPT-5 and GPT-Thinking. 

If you have used ChatGPT since August 7, you have used GPT-5. It was envisioned as being able to move back and forth between different capabilities based on the task, while representing an overall improvement in performance. “Thinking” mode performs a much deeper analysis, and shows you the steps it is going through in this process.

There have been several rounds of revisions to this rollout already, though.

ChatGPT 4o superfans vocally protested the change, finding GPT-5’s “personality” substantially less likeable, and on Monday OpenAI restored GPT 4o as a “legacy model” for paid subscribers. [Find it in your settings, under “General”, “Show additional models”] 

Wednesday, in addition to switching GPT-5 and GPT-5 Thinking to Auto, Fast, Thinking Mini, and Thinking, GPT-4.1, GPT o3, and GPT o4 mini returned. That’s (counts fingers) eight models to choose from.

Confused yet? Don’t worry about it.  Just keep using GPT-5 (your ChatGPT defaults to it)  and it’ll be fine. If you don’t like its “personality,” you can go back to 4o. Try Thinking out– it’s cool. And don’t worry too much about the rest. It’ll probably just change again anyway.

About the Author:

Sarah Pita is a fundraising professional with 25+ years of experience and a dynamic speaker who makes AI approachable and immediately useful for nonprofit teams. She leads practical, engaging trainings and workshops on using AI for fundraising and has presented at groups such as Women In Development NYC and at the AFP GPC Leading Philanthropy conference, among others. Sarah is currently Director of Development at the Center for Independence of the Disabled, New York.

Interested in an AI workshop or training? Contact Sarah here.