So, You're Not a Fan of GPT-5? Here's How to Get GPT-4o Back
Alright, let's talk about it. The new, shiny GPT-5 just rolled out, & honestly, the reception has been… mixed. For every person excited about the new features, there seems to be someone else missing the old GPT-4o. If you're in the latter camp, you're not alone. I've seen tons of posts on Reddit from people who are pretty bummed out about the change.
Turns out, a lot of users developed a real connection with GPT-4o. They loved its unique "vibe," its creativity, & the way it seemed to just get them. For some, it was more than just a tool; it was a creative partner or even a companion. The switch to GPT-5, which some are calling more "corporate" & "stiff," has been a bit of a jarring transition.
So, the big question is: can you go back? The short answer is YES, for now. OpenAI has thankfully provided a way to access what they're calling "legacy models." Here’s the thing, it's not super obvious, but I'll walk you through how to do it.
The "Legacy Models" Trick: Your Ticket Back to GPT-4o
If you're a Plus user & you're finding that GPT-5 isn't quite hitting the mark for you, don't despair. OpenAI has built-in a way to switch back to the older models. Here’s a step-by-step guide to get you back to your beloved GPT-4o.
On the Web App:
Log in to ChatGPT: Head over to the OpenAI website & log in to your account.
Find Your Profile: Look for your name or profile icon in the bottom left corner of the screen.
Go to Settings: Click on your profile, & a menu will pop up. Select "Settings."
Toggle on Legacy Models: In the "General" section of the settings, you'll see an option that says "Show legacy models." Toggle that switch on.
Check Your Model Selector: Now, when you go back to your chat window, you should see a dropdown menu at the top where you can select your model. Along with the new GPT-5 options, you should now see "legacy models" as a category, with GPT-4o nestled inside.
It's a pretty straightforward process once you know where to look. It seems like OpenAI is listening to user feedback & understands that not everyone is ready to move on from the models they've grown to love.
On the Mobile App:
If you're a mobile user, the process might be a little different. Sometimes the app can be a bit finicky when it comes to updating with the latest settings. If you've enabled legacy models on the web but you're not seeing them in the app, here are a few things to try:
Log out & log back in: Sometimes a simple refresh is all it takes.
Delete & reinstall the app: This is the go-to solution for a lot of app-related issues. Deleting the app & reinstalling it from the app store will clear the cache & force it to update with the latest settings.
Be patient: If you've tried everything & you're still not seeing the option, give it a little time. It can sometimes take an hour or two for the changes to fully propagate to your account.
Understanding the New GPT-5 Landscape
So what's the deal with GPT-5 anyway? It's not just a single new model; it's a whole new system. OpenAI is trying to create a more unified experience, where the AI automatically chooses the best way to respond to your prompt.
Here’s a quick breakdown of what you're seeing in that model selector:
GPT-5: This is the new default. It's designed to be fast & efficient for simple tasks.
GPT-5 Thinking: This is the more powerful, "thoughtful" mode. It's meant for more complex questions that require deeper reasoning. It's slower, but the results are generally better. You can either let GPT-5 switch to this mode automatically, or you can force it by adding "think hard" to your prompt.
GPT-5 Pro: This is for paid and team plans and uses more compute for the most difficult problems.
The idea is that GPT-5 will be smart enough to know when to use which mode. But as we've seen, that "auto-routing" can sometimes miss the mark. That's why having the option to manually switch back to GPT-4o is so important for a lot of users.
Why Do People Still Prefer GPT-4o?
It's a valid question. With all the hype around new models, why are so many people looking to go back? Here's what I've gathered from the online chatter:
Personality & Vibe: A lot of users felt that GPT-4o had a more distinct personality. It was more creative, more collaborative, & less formal. It felt less like a machine & more like a partner.
Continuity: For those who used ChatGPT for long-term projects or personal journaling, GPT-4o's consistent "mind" was a big deal. They had built up a history with the model, & the switch to GPT-5 felt like starting over with a stranger.
Creativity: Some users have reported that GPT-5 feels like a "massive downgrade in creativity" compared to older models. This is a huge deal for writers, artists, & anyone who relies on ChatGPT for brainstorming & creative work.
Performance: Interestingly, some offline IQ tests have shown GPT-5 scoring slightly lower than GPT-4o. While these tests aren't the be-all & end-all of model evaluation, they do add to the concern that the new model might not be an upgrade in every way.
Of course, this is all subjective. What one person finds to be a "downgrade," another might see as an improvement. The key is having the choice to use the model that works best for you.
A Note on Business & Customer Service
This whole situation highlights a really important point for businesses that are using AI. When you're interacting with customers, consistency & a personal touch are EVERYTHING. If you're using an AI chatbot for customer service or lead generation, you can't have the personality & tone of your brand changing overnight.
That's where a platform like Arsturn comes in. It helps businesses build no-code AI chatbots that are trained on their own data. This means you have complete control over the chatbot's personality, tone, & knowledge base. You can ensure that your customers are getting a consistent & personalized experience every single time they interact with your brand. You're not at the mercy of a big tech company's model updates. You can build a chatbot that truly reflects your brand's voice & values.
What About the Future?
So, what does the future hold for GPT-4o & other legacy models? Sam Altman has said that OpenAI will be monitoring the usage of GPT-4o to decide how long to keep it around. That means if you're a fan of the older model, the best thing you can do is keep using it. The more people who are actively choosing GPT-4o, the more likely it is that OpenAI will keep it as an option.
For developers, the good news is that you'll likely always have access to older models through the API & the OpenAI Playground. This gives you a lot more flexibility to choose the right model for your specific application.
In the end, it's all about having choices. The world of AI is moving at a breakneck pace, but that doesn't mean we have to leave behind the tools that we've come to rely on. Whether you're a die-hard GPT-4o fan or you're excited about what GPT-5 has to offer, it's pretty cool that we have the option to choose the AI that best fits our needs.
I hope this was helpful! Let me know what you think. Are you team GPT-4o or are you loving the new GPT-5? It'll be interesting to see how this all plays out.