OpenAI has announced that it will no longer remove old models without warning after a backlash against the introduction of GPT-5 in ChatGPT. According to Nick Turley, head of ChatGPT, the company recognizes the need for “predictability” when introducing significant changes and plans to provide advance notice before retiring any existing models.
Turley stated that the initial decision to remove GPT-4 was made to simplify model choices for ChatGPT’s 700 million weekly users. However, after a vocal part of the user base protested the change, OpenAI has reconsidered its approach. The company will now allow users to opt-in to older models, including GPT-4.
Turley emphasized that simplicity is a key goal for OpenAI and aims to make it easier for users to interact with ChatGPT without having to figure out which model to use. He also acknowledged the strong attachment people have to certain models, particularly GPT-4, which was seen as more “warm” and less annoying than GPT-5.
The change is a result of OpenAI’s efforts to balance the needs of its power users with those of its broader consumer base. Turley noted that despite the criticism, ChatGPT’s usage has increased since the release of GPT-5, highlighting the need for predictability when introducing significant changes.
Going forward, OpenAI will provide advance notice before removing any existing models and allow users to opt-in to older models. This change aims to provide a more predictable experience for users while still allowing the company to innovate and improve its products.
Source: https://www.theverge.com/openai/758537/chatgpt-4o-gpt-5-model-backlash-replacement