Retiring GPT-4o, GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini in ChatGPT
OpenAI
OpenAI is pulling GPT-4o, GPT-4.1, GPT-4.1 mini, and o4-mini from ChatGPT on Feb 13, 2026. They're also killing off GPT-5's three flavors that day — a mass model cleanup, not a one-off.
Based on reporting by OpenAI — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
OpenAI just added another wave to its ChatGPT model purge. On February 13, 2026, the company will retire GPT-4o, GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini from the ChatGPT app and web interface. This lands on the same day OpenAI already planned to sunset GPT-5 Instant, GPT-5 Thinking, and GPT-5 Pro, meaning ChatGPT users will wake up that week to a noticeably thinner model picker.
The scope here is bigger than a routine deprecation. GPT-4o was, for a long stretch, the default workhorse model for tens of millions of ChatGPT users and the face of OpenAI's multimodal push. GPT-4.1 and its mini variant were pitched as faster, cheaper successors aimed at developers and power users who wanted more control over cost and latency. o4-mini, meanwhile, was OpenAI's budget-friendly reasoning model, built for people who wanted some chain-of-thought smarts without paying full price for it. Losing all four in one sweep, right alongside the entire GPT-5 lineup, clears out most of what ChatGPT users have relied on for the past year and a half.
OpenAI's phrasing is notably narrow: this retirement applies to ChatGPT specifically. The API is untouched, at least for now, so developers building products on top of these models through OpenAI's platform don't need to scramble this month. That distinction matters. ChatGPT is where OpenAI controls the full experience and can push users toward whatever it considers the current best option — likely newer models optimized for cost, speed, or safety improvements the company hasn't detailed yet. The API, where OpenAI's revenue and reputation depend on stability for paying developers, gets a longer leash.
What OpenAI hasn't spelled out is what fills the gap. No successor names, no feature comparison, no timeline for what ChatGPT's model lineup looks like after February 13. That silence is consistent with how these retirements tend to go: announced first, replacement narrative filled in later once the older models are already gone.
My take — AI-written commentary, not fact-checked reporting
This is OpenAI doing spring cleaning on a model catalog that grew messy fast — four or five names deep with overlapping use cases nobody outside the company could fully explain. I don't mind consolidation, but the pattern of retiring first and explaining later trains users to distrust whatever model they're on, since it might vanish in a quarter. If OpenAI wants ChatGPT to feel like a stable product rather than a rotating beta, it should ship the replacement story alongside the retirement notice, not after.
Read more about this at: OpenAI