Model Retirement Overview
OpenAI confirmed plans to retire several older models on February 13, 2026, including the much-discussed GPT-4o. This decision follows the notable adoption of GPT-5.2, which now dominates usage statistics, with only 0.1% of users still opting for GPT-4o daily. The transition reflects OpenAI’s strategy to streamline its offerings and focus on the models that users actively engage with.
Why Retire GPT-4o Now?
The retirement announcement comes after a period of mixed feedback regarding GPT-4o. Initially launched in May 2024, it gained popularity for its engaging conversational style, often seen as more personable than its successors. However, GPT-5.2 has met expectations, leading OpenAI to justify the model’s retirement despite some user attachment. The company aims to improve user experience by concentrating on the most utilized models, even when it means sacrificing long-time favorites.
The Transition to GPT-5.2
GPT-5.2, released in December 2025, includes significant enhancements based on user feedback. Features like the new Personality option allow for increased customization, addressing earlier complaints about models being overly cautious. OpenAI’s focus on personalization and new safeguards aims to replicate aspects of GPT-4o’s warmth while improving functionality and reliability.
Business Implications and User Impact
For businesses leveraging OpenAI’s models, the retirement of GPT-4o may require adjustments in workflows, especially for those who relied on its unique style for creative tasks. Conversations and projects using retired models will remain accessible until the cutoff date, after which they will default to GPT-5.2. OpenAI recognizes the minimal percentage of remaining GPT-4o users but remains committed to transparency in future model deprecations.
Future Predictions
The ongoing evolution of OpenAI’s models suggests we will see continued shifts in user preferences and model capabilities. In the next 6–12 months, expect further enhancements in GPT-5.2, with potential expansions into less conventional applications, including adult content support. The focus will likely remain on balancing user engagement with operational efficiency, as OpenAI navigates the complexities of AI model evolution.
