OpenAI to Retire GPT-4 from ChatGPT as Newer Model Takes Over

OpenAI recently announced that it will begin sunsetting the original GPT-4 from its popular ChatGPT platform, beginning April 30, 2024. This model will be deprecated after the release of GPT-4 Turbo in November 2023. GPT-4 Turbo will be available to replace GPT-4 as the default model for all users. Though GPT-4 will no longer be offered through ChatGPT, it will still be available through OpenAI’s API.

GPT-4, which first appeared in March 2023, represented a big leap forward from OpenAI. This model was one of the first widely deployed multimodal models. It would be a system that knew equally well image and text. This simple fact means that GPT-4 is, indeed, a transformative innovation in artificial intelligence. It illustrates the possibilities when users and machines are able to cultivate more sophisticated dialogues.

The release of GPT-4 was not cheap, with its training costs reportedly over $100 million. Its size and complexity present it as a challenge and a powerful tool of our time spanning many applications. We recommend it for creative writing, coding projects and STEM-related challenges. MARKETPLACE pitch this year! The issue has been contentious, with major publishers including entities such as The New York Times engaged in a months-long copyright battle over the model. These lawsuits come on the heels of worries about the appropriation of copyrighted works to train AI systems.

As OpenAI moves towards this new generation of AI models, a new default model, GPT-4o, has already appeared as the default for ChatGPT. According to OpenAI, “In head-to-head evaluations, GPT-4o consistently surpasses GPT-4 in writing, coding, STEM, and more.” The company appears to be preparing to release a new improved family of models named GPT-4.1. This transition is indicative of their deep commitment to advancing innovation and enriching their AI products.

GPT-4 as the underlying model for direct consumer use via ChatGPT will no longer be immediately available. Developers and organizations can continue to harness this powerful model through the API. This decision reflects OpenAI’s commitment to keeping our technology open and allowing a wide range of applications and innovative solutions with our technology.