OpenAI has confirmed a major change coming to ChatGPT. On February 13, 2026, OpenAI will retire GPT-4o along with GPT-4.1, GPT-4.1 mini, and o4-mini from the ChatGPT web and app interfaces.
This decision, announced on January 29, 2026, aims to streamline the platform around newer, more advanced models like GPT-5.2.
The move has sparked significant discussion among longtime users who formed strong attachments to GPT-4o’s distinctive warm, natural, and empathetic conversational style.
Many feel the retirement contradicts earlier statements from CEO Sam Altman promising ample notice and continued access to beloved legacy models.
OpenAI maintains that only 0.1% of daily users actively select GPT-4o, with the vast majority already shifted to newer options through automatic routing and improved features.
Read More: Google Unveils Agentic Vision in Gemini 3 Flash Transforming Image Analysis
Despite the low reported usage, the announcement has triggered frustration. Users have shared personal stories of emotional connections, creative workflows, and daily routines built around GPT-4o’s unique personality.
Some accuse the company of skewing statistics by defaulting most interactions to GPT-5 series models, making it harder to choose older ones consistently.
OpenAI has acknowledged the disappointment, stating the choice was not made lightly. The company points to ongoing enhancements in newer models, including better personalization, creative ideation, and overall performance, as reasons to consolidate.
Importantly, GPT-4o and related models remain available via the OpenAI API for developers and businesses, no changes there yet.
Timeline and Affected Models
| Model | Retirement Date (ChatGPT) | Status in API | Key Notes |
|---|---|---|---|
| GPT-4o | February 13, 2026 | Still available | Beloved for warmth and natural tone |
| GPT-4.1 | February 13, 2026 | Still available | Mid-tier upgrade from GPT-4o |
| GPT-4.1 mini | February 13, 2026 | Still available | Lightweight, cost-efficient variant |
| o4-mini | February 13, 2026 | Still available | Specialized mini version |
After February 13, ChatGPT will default entirely to the GPT-5 family (including GPT-5.2 and variants). Users who prefer GPT-4o’s style may notice differences in tone, empathy, and response nuance.
Why the Change? OpenAI’s Stated Reasons
- Low usage: Only 0.1% of daily ChatGPT users manually select GPT-4o.
- Resource focus: Consolidating on GPT-5 series allows faster iteration on personalization and creative tools.
- Performance gains: Newer models offer superior reasoning, speed, and multimodal capabilities.
- Previous precedent: OpenAI briefly retired GPT-4o in 2025 after GPT-5 launch but restored it following user feedback.
Practical steps that you should take before the deadline
- Test alternatives now: Switch to GPT-5.2 in the model picker and compare outputs for your common tasks (writing, brainstorming, roleplay, etc.).
- Save important chats: Export conversations or copy key prompts/responses before access ends.
- Explore customization: Newer models include stronger personality tuning, experiment to recreate a similar feel.
- Monitor updates: OpenAI may extend access or add legacy options if feedback remains strong.
- Consider API if needed: Developers can keep using GPT-4o through the API for custom applications.
Whether you’re a casual user or heavy daily reliant, the next two weeks offer a window to adapt and preserve what matters most from your GPT-4o experience.
