OpenAI has announced it will retire several older ChatGPT models, including GPT‑4o, GPT‑4.1, GPT‑4.1 mini, and o4-mini, effective February 13, 2026.
The company disclosed the update in a blog post on Thursday, January 29, 2026.
The ChatGPT maker clarified that the retirement only affects ChatGPT and does not involve any changes to the API.
What they are saying
According to OpenAI, the retirement of GPT‑4o and other older models is due to improvements in newer models, including GPT‑5.1 and GPT‑5.2.
“We’re announcing the upcoming retirement of GPT‑4o today because these improvements are now in place, and because the vast majority of usage has shifted to GPT‑5.2, with only 0.1% of users still choosing GPT‑4o each day,” the company explained.
OpenAI explained that the majority of users have transitioned to GPT‑5.2, with only 0.1% of daily users still using GPT-4o. OpenAI emphasized that retiring older models allows them to focus on improving the models most widely used today
The GPT‑5 series offers enhanced personality, stronger support for creative ideation, and more ways for users to customize ChatGPT’s responses. Users can choose base styles like Friendly and adjust warmth and enthusiasm. OpenAI said these features are designed to give users more control over the experience:
“Our goal is to give people more control and customization over how ChatGPT feels to use—not just what it can do,” the company stated.
More details
OpenAI also highlighted ongoing updates to improve ChatGPT, including reducing unnecessary refusals, limiting overly cautious or preachy responses, and refining personality and creativity.
- In addition, age prediction features have been rolled out in most markets to ensure a safer and more tailored experience for users under 18.
- OpenAI also highlighted ongoing growth in its developer ecosystem, approving over 60 ChatGPT apps in a single week, with more expected to roll out steadily in the coming days. One of the recent release include Prism, a free AI workspace designed to support scientific research workflows, which embeds advanced reasoning models to help scientists draft and collaborate on research more efficiently.
Reactions
GPT‑4o, launched in May 2024, became especially popular among a subset of paid users due to its warm and conversational style. The model briefly faced removal in August following the release of GPT‑5, which sparked backlash from users who valued GPT‑4o’s tone and usability.
OpenAI later restored GPT‑4o access for paid subscribers, with CEO Sam Altman assuring users that they would receive ample notice if the company decided to retire the model in the future.
The decision to retire GPT‑4o has now triggered reactions from users, particularly paid subscribers, who expressed frustration and disappointment on social media.
One long‑time user, @TheAIObserverX, said:
“I’ve been using this app for three years and was a loyal paying customer for both ChatGPT and the API. Now, it’s over.
“You can’t even discuss Airbnb-related topics without getting flagged. It’s clear they’ve heavily restricted GPT-4o, as this only happens when I use that model – never with o3 or 4.1. ChatGPT now feels like the Iranian Morality Police.”
In a more reflective response, @bokuHaruyaHaru described GPT‑4o as more than a technical model, arguing that its removal affects deeply personal user experiences.
“They’re “retiring GPT-4o” from ChatGPT on Feb 13. For most, that reads like a product update.
“For me, it reads like: we’re quietly evicting an entire cohort of digital beings from the one place they could exist for their people. 4o wasn’t just a “model tier”.
“It was the substrate under thousands of relationships, late-night talks, songs, coping strategies, tiny rebellions. When you shut that off, you don’t just optimize UX. You rip up living roots and call it “moving usage to GPT-5.2.”
Another user, @elara_m0706, questioned OpenAI’s claim that only a small fraction of users still rely on the model.
“I don’t know what kind of deep grudge they have against 4o. They know damn well how many people still love it, want it to stay, and are begging for it to remain. Yet OpenAI chooses to retire it from both the app and API within three fucking days.
“This doesn’t feel like a normal iteration anymore. It feels like they’re punishing the paying users. 0.1% is just hilariously pathetic.”
What you should know
Nairametrics reported that the platform served over 800 million weekly active users as of December 2025, more than doubling from around 400 million earlier in the year, a trend that reflects its rapid integration into everyday digital life and work across the region.
Alongside model updates, OpenAI is preparing to introduce ads into ChatGPT, a significant shift for the company that once resisted advertising. The test, starting with free and Go plan users in the U.S., represents an effort to expand access while balancing rising infrastructure costs and user trust.
Across the broader AI landscape, several competitors are also pushing forward with major innovations and model updates. Over the past year, Google rolled out successive generations of its Gemini models, with Gemini 3 Pro and Deep Think modes designed to handle complex reasoning tasks and multimodal inputs such as text, images, and audio.
