Microsoft-backed OpenAI announced that it will retire the GPT-4 model from ChatGPT on April 30, as noted in a changelog update released Thursday. The company will fully transition to GPT-4o, which is now the default model on the platform. “Effective April 30, GPT-4 will be fully replaced by GPT-4o,” the company stated. While GPT-4 will no longer be available within ChatGPT, it will remain accessible to developers through OpenAI’s API.
According to OpenAI, GPT-4o offers superior performance across a variety of tasks. “In head-to-head evaluations, GPT-4o consistently outperforms GPT-4 in writing, coding, STEM subjects, and more,” the company said. “Recent enhancements have also improved its ability to follow instructions, solve problems, and maintain conversational flow – positioning it as a natural successor to GPT-4.”
GPT-4, launched in March 2023, was OpenAI’s first widely adopted multimodal model, capable of processing both text and images. It powered ChatGPT and Microsoft Copilot. OpenAI CEO Sam Altman previously revealed that training GPT-4 cost over $100 million.
The model has also been at the center of ongoing copyright lawsuits, including one filed by The New York Times. The publication claims OpenAI used its content to train GPT-4 without authorization. OpenAI has denied any wrongdoing, asserting that its use of publicly available data falls under fair use.
In parallel with the model transition, OpenAI has rolled out a new memory feature in ChatGPT. This enhancement enables the chatbot to recall and reference previous conversations, making its responses more personalized and contextually aware.
“We’ve significantly upgraded memory in ChatGPT – it can now remember all your past conversations!” Altman shared in a post on X (formerly Twitter).
