OpenAI has revealed its plan to retire the GPT-4 language model from ChatGPT by April 30.
It will be fully replaced by the newer GPT-4o, which is now the default model on the platform, according to the official release notes.
While it will no longer be available inside ChatGPT, GPT-4 will remain accessible through the company’s API.
According to OpenAI’s changelog, GPT-4o significantly outperforms GPT-4 in various areas, including writing, coding, and technical reasoning.
OpenAI CEO Sam Altman previously stated that GPT-4’s training cost exceeded $100 million, reflecting the scale and complexity of the model.
GPT-4 was first introduced in March 2023, becoming part of both ChatGPT and Microsoft Copilot.
It was known for its multi-modal capabilities, allowing it to process both text and images simultaneously—something that marked a turning point in OpenAI’s widely deployed models.
OpenAI CEO Sam Altman previously stated that GPT-4’s training cost exceeded $100 million, reflecting the scale and complexity of the model.
In November 2023, the company launched a faster, more cost-effective variant known as GPT-4 Turbo, which was seen as a more efficient alternative.
In addition to the upcoming deactivation of GPT-4, tech sources report that OpenAI is preparing to launch a new series of models under the name GPT-4.1. These will include versions such as GPT-4.1-mini and GPT-4.1-nano.
The company is also planning to release a new inference model named o4-mini, along with the main version of o3.
