OpenAI to Phase Out GPT-4o and Older Models from ChatGPT, Gives Users Two-Week Transition Window
OpenAI will retire GPT-4o and several legacy ChatGPT models next month, giving users two weeks to transition before removal on February 13, 2026. The company says the change will let it concentrate on faster, more capable systems that now handle the bulk of day-to-day tasks.
What’s changing
OpenAI confirmed it will remove the following models from ChatGPT:
- GPT-4o
- GPT-5 (Instant and Thinking)
- GPT-4.1
- GPT-4.1 mini
- OpenAI o4-mini
For developers, OpenAI says there are no immediate changes to its APIs, so existing integrations can continue to operate for now.
Why GPT-4o mattered
Launched in May 2024, GPT-4o quickly earned a following for its warm tone and natural, personable conversations. When OpenAI temporarily pulled the model after introducing GPT-5, pushback from loyal users was swift. The company reversed course at the time, underscoring how central GPT-4o’s “voice” had become to creative and brainstorming workflows.
“We brought GPT4o back after hearing clear feedback from a subset of Plus and Pro users, who told us they needed more time to transition key use cases, like creative ideation, and that they preferred GPT4o’s conversational style and warmth.”
Usage has shifted to newer defaults
As OpenAI’s latest-generation systems improved, user behavior evolved. According to the company’s internal metrics, only about 0.1% of users select GPT-4o on a typical day, while GPT-5.2 has become the default for most daily tasks. With that shift, OpenAI says it needs to streamline its lineup and focus on the models that serve the majority.
“Retiring models is never easy, but it allows us to focus on improving the models most people use today.”
Preserving GPT-4o’s personality
To soften the transition, OpenAI has added customization features designed to capture what users loved about GPT-4o. After last year’s controversy, the company rolled out adjustable base styles and tones—such as “Friendly”—and controls for warmth and enthusiasm. These tools make it easier to tune newer models so they feel more conversational and approachable.
Two-week window and what to do now
OpenAI had previously promised advance notice before phasing out older models, and this announcement honors that commitment. With the February 13 cutoff approaching, users should take a few practical steps:
- Identify any chats or workflows that rely on GPT-4o’s specific tone or behavior and migrate them to newer models.
- Use style presets (e.g., Friendly) and adjust warmth/enthusiasm to approximate GPT-4o’s voice.
- Test GPT-5.2 for daily tasks and compare performance on creative ideation, brainstorming, and conversational use cases.
- Export or archive important conversations that reference GPT-4o behavior for future reference.
- If you build on OpenAI APIs, note that nothing changes immediately—but keep an eye on future deprecation timelines.
The bigger picture
While some users will miss GPT-4o’s familiar presence, the move reflects OpenAI’s broader push toward faster, smarter, and more efficient AI systems. As the company narrows its focus on next-generation models, ChatGPT users should see continued upgrades—though it comes with a goodbye to one of the platform’s most personable models.
