OpenAI has announced that it will officially retire its GPT-4 model in ChatGPT by April 30, 2025. The company is transitioning to GPT-4o, a more advanced version offering significant improvements in coding, science, and mathematics.
The Microsoft-backed AI firm revealed on Thursday that GPT-4o consistently outperforms GPT-4 in instruction-following and conversational tasks, thanks to recent updates that enhanced its performance and reliability.
With over 400 million weekly active users, ChatGPT users can now access a more responsive and capable AI, designed to deliver cleaner front-end code and tackle complex technical problems with greater accuracy.
A statement from OpenAI highlighted the upgrade:
*”Recent improvements have further refined GPT‑4o’s instruction following, problem-solving skills, and natural conversational flow, making it the clear successor to GPT-4.”*
Additionally, the AI model now excels at generating clean, efficient code, accurately reviewing and updating existing code, and producing outputs that compile and run successfully—streamlining developers’ workflows.
OpenAI also noted that users testing GPT-4o have reported enhanced performance in creative tasks and clearer, more concise responses in conversations.
GPT-4o is currently available as the default model in ChatGPT and through the API under the label “chatgpt-4o-latest.” The company plans to extend these improvements to earlier models available via API in the coming weeks.
