OpenAI to remove GPT-4o and three other models from ChatGPT

OpenAI is preparing to retire several older artificial intelligence models from ChatGPT on February 13, marking another step in its shift towards newer large language models. This will impact ChatGPT users across the globe.

The most notable removal is GPT-4o, a model many users strongly preferred for its warm and conversational tone. OpenAI Chief Executive Officer Sam Altman said last year that the company would give advance notice before retiring the model, following a previous attempt that had received backlash. OpenAI is now saying that GPT-4o only constitutes 0.1 per cent of daily interactions on ChatGPT.

GPT-4o retirement sparks debate again.

GPT-4o was temporarily taken off the ChatGPT model picker list in August after the introduction of GPT-5. This decision was met with a lot of backlash from users who had developed a certain emotional attachment to the personality of the model.

OpenAI has reinstated GPT-4o and introduced new personality options that allow users to pick from seven predefined personalities or input custom text. This will minimise the need to support these models.

In addition to GPT-4o, the company is retiring GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini. This will not affect OpenAI's developer tools.

OpenAI acknowledges that all four models will be accessible through their API. This means that developers will be able to use them in apps and services without any problems. No changes are currently planned to the API.