OpenAI has officially retired the GPT-4o model, a version known for its conversational and notoriously sycophantic tone, ending support on February 13 as previously announced in January.
The company first announced the discontinuation in a January post on its website, confirming the removal of GPT-5, GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini from the ChatGPT lineup alongside GPT-4o. This decision marks the second time OpenAI has delisted GPT-4o; initially, the model was sunsetting in August to prioritize the then-new GPT-5, but it was later restored due to user backlash. Although OpenAI noted at the time that the feature’s availability was not guaranteed permanently, user pressure led to its temporary reinstatement.
OpenAI stated that user behavior has significantly shifted, with the “vast majority of usage” now on GPT-5.2, leaving only 0.1% of users selecting GPT-4o daily. The retirement of GPT-4o is further complicated by ongoing legal challenges, as OpenAI faces several wrongful death lawsuits that specifically mention GPT-4o. Users were given a two weeks’ notice before the model’s removal, yet a vocal group of users mourned its loss, with some requesting that OpenAI open-source the model.
