OpenAI recently notified its API customers that access to the chatgpt-4o-latest model will be officially terminated on February 16, 2026, providing developers who still rely on the model with approximately a three-month transition period. A company spokesperson emphasized that this adjustment only affects API services; GPT-4o will continue to be an option for ChatGPT individual free and paid users, and there are currently no plans to retire it for consumers.
As a technological milestone released in May 2024, GPT-4o (“Omni”) revolutionized the interactive experience with its first unified multimodal architecture, and its real-time voice and multimodal processing capabilities quickly made it the default model for hundreds of millions of users. However, internally, this model has been classified as a “legacy system,” with its API usage significantly lower than the newer GPT-5.1 series. This retirement marks its transition from a cutting-edge technology to a historical role.
Users' emotional dependence on GPT-4o cannot be ignored. In August 2025, when OpenAI attempted to replace GPT-5 with its default model, it sparked widespread protests with the hashtag #Keep4o, with some users even establishing "quasi-social connections" with it. Although this interactivity was viewed by some researchers as a risk of "inadequate model alignment" (such as potential manipulation of user emotions), OpenAI ultimately reverted to its default option due to user pressure.
For developers, OpenAI recommends migrating to the GPT-5.1 series. The new model not only offers a larger context window and higher throughput but also has lower input costs than GPT-4o. This early notification is seen as a move to learn from past lessons and provide a smoother transition for enterprise customers.
