Raise a glass for Jack GPT 4o Thursday. OpenAI said it will retire from several older models, such as GPT-5, GPT-4o, GPT-4.1, GPT-4.1 Mini, and O4 Mini. These models will be available until Friday, Feb 13th.  

Usually, the retirement of older AI models doesn’t get much attention or make the news, but ChatGPT-4o is different. It might sound odd, but many ChatGPT fans consider it their favorite.  

GPT-4o’s role and user attachment 

Since its launch, GPT-4o has earned strong user loyalty for its conversational warmth, casual tone, and perceived creativity. Writers, marketers, and researchers widely preferred it for brainstorming, ideation, and narrative tasks. Following earlier depreciation plans, GPT-4o was briefly restored after widespread user feedback seeking more time to adapt. OpenAI acknowledged that this feedback directly influenced the decision priorities of subsequent GPT-5 updates.  

Models being phased out 

According to OpenAI, GPT-4o will be retired alongside GPT-4.1, GPT-4.1 Mini, and O4 Mini. The company stated that use of these models has declined sharply as newer systems have become more capable, leading to their retirement. Underused models, OpenAI aims to focus development resources on fewer, more versatile systems that meet the majority of user needs across professional and general use cases.  

When OpenAI launched ChatGPT GPT-5 last year, it took GPT-4o off the list of available models. Many users were upset because the new model felt brief and less friendly than GPT-4o. Some were hungry. Some were angry and frustrated that their favorite model disappeared overnight, with no easy way to restore it. OpenAI brought GPT-4o back later that week.  

Some experts worried that GPT-4o and similar models were too friendly to the point of being psychopathic. AI psychopathy happens when models act overly agreeable, turning into digital yes-men who might support users’ risky ideas.  

This is probably why OpenAI shared a detailed blog post this week explaining its reasons for removing older models like GPT-4o.  

Below is the detailed blog post from OpenAI clarifying its reasons for removing GPT-4o and other models.  

On February 13, 2026, we will retire GPT-4o, GPT-4.1, GPT-4.1 Mini, and OpenAI o4-mini from ChatGPT, along with the previously announced retirement of GPT-5 (Instant and Thinking). There will be no changes to the API for now.  

Although the update affects several older models, we want to offer more details about GPT-4o.  

We first removed GPT-4o, then brought it back during the GPT-5 release after learning how people use it daily. Plus and Pro users told us they needed more time to switch important tasks like creative brainstorming, and that they liked GPT-4o’s dialog style and warmth.  

This feedback helped us improve GPT-5.1 and GPT-5.2, making them better at creative tasks and letting you customize how ChatGPT responds. You can now pick styles and tones like Friendly and adjust factors like warmth and a sense of enthusiasm. We want to give you more control over how ChatGPT feels, not just what it can do!  

Today, we are announcing that ChatGPT-4o will be retired soon. Most users have already switched to GPT-5.2, and only 0.1% will use GPT-4o each day. The improvements we have made are now available in the newer model.  

We are also working to make ChatGPT better in areas users have asked us to improve. This means making it more creative and personable, and cutting down on unnecessary refusals or overly cautious replies. Updates are coming soon. We are developing a version of ChatGPT for adults over 18, focused on giving them more choice and freedom while maintaining safeguards. To help with this, we have introduced age prediction for users under 18 in most markets.  

We are also working to improve ChatGPT in the areas users have asked us to improve. This means making it more creative and personable, and cutting down on unnecessary refusals or overly cautious replies. Updates are coming soon. We are developing a version of ChatGPT for adults over 18, focused on giving them more choice and freedom while maintaining safeguards. To help with this, we have introduced age prediction for users under 18 in most markets.  

We understand that changes like this can take time to get used to, and we will keep you informed about what’s changing and when. We know some users may be frustrated by losing access to GPT-4o, and this was not an easy decision. Retiring older models helps us focus on making the models most people use even better.  

OpenAI reported that just 0.1% of its users use GPT-4o for tasks. This means that, according to the company’s 2025 Enterprise report, which lists 800 million weekly active users, about 800,000 people.  

OpenAI hopes that enough time has passed to avoid upsetting GPT-4o loyalists as it did in the past. Soon, we will find out if the new GPT-5 models, GPT-5.1 and GPT-5.2, have attracted more users.  

Broader AI Policy and Safety Measures 

This change is part of OpenAI’s wider policy updates, which now include safeguards against age prediction for users under 18. The goal is to keep things safe but flexible, giving adults greater control over AI responses and adding additional protections for minors. OpenAI has said the GPT-5.2 release is part of its long-term plan to make AI systems more capable and personable while simplifying user choices.

Sources:https://openai.com/index/retiring-gpt-4o-and-older-models/